From 02905a049f5bb8b102507e3de92a9915411fce9c Mon Sep 17 00:00:00 2001 From: Mateusz Sluszniak Date: Mon, 9 Jun 2025 19:01:47 +0200 Subject: [PATCH 01/20] Refactor model URL constants --- .cspell-wordlist.txt | 2 + .../src/constants/modelUrls.ts | 241 +++++++----------- 2 files changed, 92 insertions(+), 151 deletions(-) diff --git a/.cspell-wordlist.txt b/.cspell-wordlist.txt index 3742847485..03c32ae473 100644 --- a/.cspell-wordlist.txt +++ b/.cspell-wordlist.txt @@ -53,3 +53,5 @@ QINT FNUZ wordlist jitpack +coreml +mobilenetv diff --git a/packages/react-native-executorch/src/constants/modelUrls.ts b/packages/react-native-executorch/src/constants/modelUrls.ts index 71730b1857..e68ebebbce 100644 --- a/packages/react-native-executorch/src/constants/modelUrls.ts +++ b/packages/react-native-executorch/src/constants/modelUrls.ts @@ -1,191 +1,129 @@ import { Platform } from 'react-native'; -// LLM's +const BASE_URL_PREFIX = + 'https://huggingface.co/software-mansion/react-native-executorch'; +const VERSION_TAG = 'resolve/v0.4.0'; + +// LLM`s // LLAMA 3.2 -export const LLAMA3_2_3B = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-3B/original/llama3_2_3B_bf16.pte'; -export const LLAMA3_2_3B_QLORA = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-3B/QLoRA/llama3_2-3B_qat_lora.pte'; -export const LLAMA3_2_3B_SPINQUANT = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-3B/spinquant/llama3_2_3B_spinquant.pte'; -export const LLAMA3_2_1B = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-1B/original/llama3_2_bf16.pte'; -export const LLAMA3_2_1B_QLORA = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-1B/QLoRA/llama3_2_qat_lora.pte'; -export const LLAMA3_2_1B_SPINQUANT = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-1B/spinquant/llama3_2_spinquant.pte'; -export const LLAMA3_2_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/tokenizer.json'; -export const LLAMA3_2_TOKENIZER_CONFIG = - 'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/tokenizer_config.json'; +export const LLAMA3_2_3B = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/original/llama3_2_3B_bf16.pte`; +export const LLAMA3_2_3B_QLORA = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/QLoRA/llama3_2-3B_qat_lora.pte`; +export const LLAMA3_2_3B_SPINQUANT = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/spinquant/llama3_2_3B_spinquant.pte`; +export const LLAMA3_2_1B = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/original/llama3_2_bf16.pte`; +export const LLAMA3_2_1B_QLORA = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/QLoRA/llama3_2_qat_lora.pte`; +export const LLAMA3_2_1B_SPINQUANT = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/spinquant/llama3_2_spinquant.pte`; +export const LLAMA3_2_TOKENIZER = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/tokenizer.json`; +export const LLAMA3_2_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/tokenizer_config.json`; // QWEN 3 -export const QWEN3_0_6B = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-0.6B/original/qwen3_0_6b_bf16.pte'; -export const QWEN3_0_6B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-0.6B/quantized/qwen3_0_6b_8da4w.pte'; -export const QWEN3_1_7B = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-1.7B/original/qwen3_1_7b_bf16.pte'; -export const QWEN3_1_7B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-1.7B/quantized/qwen3_1_7b_8da4w.pte'; -export const QWEN3_4B = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-4B/original/qwen3_4b_bf16.pte'; -export const QWEN3_4B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-4B/quantized/qwen3_4b_8da4w.pte'; -export const QWEN3_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/tokenizer.json'; -export const QWEN3_TOKENIZER_CONFIG = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/tokenizer_config.json'; +export const QWEN3_0_6B = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-0.6B/original/qwen3_0_6b_bf16.pte`; +export const QWEN3_0_6B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-0.6B/quantized/qwen3_0_6b_8da4w.pte`; +export const QWEN3_1_7B = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-1.7B/original/qwen3_1_7b_bf16.pte`; +export const QWEN3_1_7B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-1.7B/quantized/qwen3_1_7b_8da4w.pte`; +export const QWEN3_4B = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-4B/original/qwen3_4b_bf16.pte`; +export const QWEN3_4B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-4B/quantized/qwen3_4b_8da4w.pte`; +export const QWEN3_TOKENIZER = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/tokenizer.json`; +export const QWEN3_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/tokenizer_config.json`; // HAMMER 2.1 -export const HAMMER2_1_0_5B = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-0.5B/original/hammer2_1_0_5B_bf16.pte'; -export const HAMMER2_1_0_5B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-0.5B/quantized/hammer2_1_0_5B_8da4w.pte'; -export const HAMMER2_1_1_5B = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-1.5B/original/hammer2_1_1_5B_bf16.pte'; -export const HAMMER2_1_1_5B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-1.5B/quantized/hammer2_1_1_5B_8da4w.pte'; -export const HAMMER2_1_3B = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-3B/original/hammer2_1_3B_bf16.pte'; -export const HAMMER2_1_3B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-3B/quantized/hammer2_1_3B_8da4w.pte'; -export const HAMMER2_1_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/tokenizer.json'; -export const HAMMER2_1_TOKENIZER_CONFIG = - 'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/tokenizer_config.json'; +export const HAMMER2_1_0_5B = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-0.5B/original/hammer2_1_0_5B_bf16.pte`; +export const HAMMER2_1_0_5B_QUANTIZED = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-0.5B/quantized/hammer2_1_0_5B_8da4w.pte`; +export const HAMMER2_1_1_5B = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-1.5B/original/hammer2_1_1_5B_bf16.pte`; +export const HAMMER2_1_1_5B_QUANTIZED = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-1.5B/quantized/hammer2_1_1_5B_8da4w.pte`; +export const HAMMER2_1_3B = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-3B/original/hammer2_1_3B_bf16.pte`; +export const HAMMER2_1_3B_QUANTIZED = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-3B/quantized/hammer2_1_3B_8da4w.pte`; +export const HAMMER2_1_TOKENIZER = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/tokenizer.json`; +export const HAMMER2_1_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/tokenizer_config.json`; // SMOLLM2 -export const SMOLLM2_1_135M = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-135M/original/smolLm2_135M_bf16.pte'; -export const SMOLLM2_1_135M_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-135M/quantized/smolLm2_135M_8da4w.pte'; -export const SMOLLM2_1_360M = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-360M/original/smolLm2_360M_bf16.pte'; -export const SMOLLM2_1_360M_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-360M/quantized/smolLm2_360M_8da4w.pte'; -export const SMOLLM2_1_1_7B = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-1.7B/original/smolLm2_1_7B_bf16.pte'; -export const SMOLLM2_1_1_7B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-1.7B/quantized/smolLm2_1_7B_8da4w.pte'; -export const SMOLLM2_1_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/tokenizer.json'; -export const SMOLLM2_1_TOKENIZER_CONFIG = - 'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/tokenizer_config.json'; +export const SMOLLM2_1_135M = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-135M/original/smolLm2_135M_bf16.pte`; +export const SMOLLM2_1_135M_QUANTIZED = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-135M/quantized/smolLm2_135M_8da4w.pte`; +export const SMOLLM2_1_360M = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-360M/original/smolLm2_360M_bf16.pte`; +export const SMOLLM2_1_360M_QUANTIZED = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-360M/quantized/smolLm2_360M_8da4w.pte`; +export const SMOLLM2_1_1_7B = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-1.7B/original/smolLm2_1_7B_bf16.pte`; +export const SMOLLM2_1_1_7B_QUANTIZED = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-1.7B/quantized/smolLm2_1_7B_8da4w.pte`; +export const SMOLLM2_1_TOKENIZER = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/tokenizer.json`; +export const SMOLLM2_1_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/tokenizer_config.json`; // QWEN 2.5 -export const QWEN2_5_0_5B = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-0.5B/original/qwen2_5_0_5b_bf16.pte'; -export const QWEN2_5_0_5B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-0.5B/quantized/qwen2_5_0_5b_8da4w.pte'; -export const QWEN2_5_1_5B = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-1.5B/original/qwen2_5_1_5b_bf16.pte'; -export const QWEN2_5_1_5B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-1.5B/quantized/qwen2_5_1_5b_8da4w.pte'; -export const QWEN2_5_3B = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-3B/original/qwen2_5_3b_bf16.pte'; -export const QWEN2_5_3B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-3B/quantized/qwen2_5_3b_8da4w.pte'; -export const QWEN2_5_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/tokenizer.json'; -export const QWEN2_5_TOKENIZER_CONFIG = - 'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/tokenizer_config.json'; +export const QWEN2_5_0_5B = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-0.5B/original/qwen2_5_0_5b_bf16.pte`; +export const QWEN2_5_0_5B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-0.5B/quantized/qwen2_5_0_5b_8da4w.pte`; +export const QWEN2_5_1_5B = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-1.5B/original/qwen2_5_1_5b_bf16.pte`; +export const QWEN2_5_1_5B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-1.5B/quantized/qwen2_5_1_5b_8da4w.pte`; +export const QWEN2_5_3B = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-3B/original/qwen2_5_3b_bf16.pte`; +export const QWEN2_5_3B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-3B/quantized/qwen2_5_3b_8da4w.pte`; +export const QWEN2_5_TOKENIZER = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/tokenizer.json`; +export const QWEN2_5_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/tokenizer_config.json`; // PHI 4 -export const PHI_4_MINI_4B = - 'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/original/phi-4-mini_bf16.pte'; -export const PHI_4_MINI_4B_QUANTIZED = - 'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/quantized/phi-4-mini_8da4w.pte'; -export const PHI_4_MINI_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/tokenizer.json'; -export const PHI_4_MINI_TOKENIZER_CONFIG = - 'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/tokenizer_config.json'; +export const PHI_4_MINI_4B = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/original/phi-4-mini_bf16.pte`; +export const PHI_4_MINI_4B_QUANTIZED = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/quantized/phi-4-mini_8da4w.pte`; +export const PHI_4_MINI_TOKENIZER = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/tokenizer.json`; +export const PHI_4_MINI_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/tokenizer_config.json`; // Classification export const EFFICIENTNET_V2_S = - Platform.OS === 'ios' - ? 'https://huggingface.co/software-mansion/react-native-executorch-efficientnet-v2-s/resolve/v0.4.0/coreml/efficientnet_v2_s_coreml_all.pte' - : 'https://huggingface.co/software-mansion/react-native-executorch-efficientnet-v2-s/resolve/v0.4.0/xnnpack/efficientnet_v2_s_xnnpack.pte'; + Platform.OS === `ios` + ? `${BASE_URL_PREFIX}-efficientnet-v2-s/${VERSION_TAG}/coreml/efficientnet_v2_s_coreml_all.pte` + : `${BASE_URL_PREFIX}-efficientnet-v2-s/${VERSION_TAG}/xnnpack/efficientnet_v2_s_xnnpack.pte`; // Object detection -export const SSDLITE_320_MOBILENET_V3_LARGE = - 'https://huggingface.co/software-mansion/react-native-executorch-ssdlite320-mobilenet-v3-large/resolve/v0.4.0/ssdlite320-mobilenetv3-large.pte'; +export const SSDLITE_320_MOBILENET_V3_LARGE = `${BASE_URL_PREFIX}-ssdlite320-mobilenet-v3-large/${VERSION_TAG}/ssdlite320-mobilenetv3-large.pte`; // Style transfer export const STYLE_TRANSFER_CANDY = - Platform.OS === 'ios' - ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-candy/resolve/v0.4.0/coreml/style_transfer_candy_coreml.pte' - : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-candy/resolve/v0.4.0/xnnpack/style_transfer_candy_xnnpack.pte'; + Platform.OS === `ios` + ? `${BASE_URL_PREFIX}-style-transfer-candy/${VERSION_TAG}/coreml/style_transfer_candy_coreml.pte` + : `${BASE_URL_PREFIX}-style-transfer-candy/${VERSION_TAG}/xnnpack/style_transfer_candy_xnnpack.pte`; export const STYLE_TRANSFER_MOSAIC = - Platform.OS === 'ios' - ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-mosaic/resolve/v0.4.0/coreml/style_transfer_mosaic_coreml.pte' - : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-mosaic/resolve/v0.4.0/xnnpack/style_transfer_mosaic_xnnpack.pte'; + Platform.OS === `ios` + ? `${BASE_URL_PREFIX}-style-transfer-mosaic/${VERSION_TAG}/coreml/style_transfer_mosaic_coreml.pte` + : `${BASE_URL_PREFIX}-style-transfer-mosaic/${VERSION_TAG}/xnnpack/style_transfer_mosaic_xnnpack.pte`; export const STYLE_TRANSFER_RAIN_PRINCESS = - Platform.OS === 'ios' - ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-rain-princess/resolve/v0.4.0/coreml/style_transfer_rain_princess_coreml.pte' - : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-rain-princess/resolve/v0.4.0/xnnpack/style_transfer_rain_princess_xnnpack.pte'; + Platform.OS === `ios` + ? `${BASE_URL_PREFIX}-style-transfer-rain-princess/${VERSION_TAG}/coreml/style_transfer_rain_princess_coreml.pte` + : `${BASE_URL_PREFIX}-style-transfer-rain-princess/${VERSION_TAG}/xnnpack/style_transfer_rain_princess_xnnpack.pte`; export const STYLE_TRANSFER_UDNIE = - Platform.OS === 'ios' - ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-udnie/resolve/v0.4.0/coreml/style_transfer_udnie_coreml.pte' - : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-udnie/resolve/v0.4.0/xnnpack/style_transfer_udnie_xnnpack.pte'; + Platform.OS === `ios` + ? `${BASE_URL_PREFIX}-style-transfer-udnie/${VERSION_TAG}/coreml/style_transfer_udnie_coreml.pte` + : `${BASE_URL_PREFIX}-style-transfer-udnie/${VERSION_TAG}/xnnpack/style_transfer_udnie_xnnpack.pte`; // S2T -export const MOONSHINE_TINY_DECODER = - 'https://huggingface.co/software-mansion/react-native-executorch-moonshine-tiny/resolve/v0.4.0/xnnpack/moonshine_tiny_xnnpack_decoder.pte'; -export const MOONSHINE_TINY_ENCODER = - 'https://huggingface.co/software-mansion/react-native-executorch-moonshine-tiny/resolve/v0.4.0/xnnpack/moonshine_tiny_xnnpack_encoder.pte'; -export const MOONSHINE_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-moonshine-tiny/resolve/v0.4.0/moonshine_tiny_tokenizer.json'; -export const WHISPER_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny.en/resolve/v0.4.0/whisper_tokenizer.json'; -export const WHISPER_TINY_DECODER = - 'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny.en/resolve/v0.4.0/xnnpack/whisper_tiny_en_xnnpack_decoder.pte'; -export const WHISPER_TINY_ENCODER = - 'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny.en/resolve/v0.4.0/xnnpack/whisper_tiny_en_xnnpack_encoder.pte'; -export const WHISPER_TINY_MULTILINGUAL_ENCODER = - 'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny/resolve/v0.4.0/xnnpack/xnnpack_whisper_encoder.pte'; -export const WHISPER_TINY_MULTILINGUAL_DECODER = - 'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny/resolve/v0.4.0/xnnpack/xnnpack_whisper_decoder.pte'; -export const WHISPER_TINY_MULTILINGUAL_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny/resolve/v0.4.0/tokenizer.json'; +export const MOONSHINE_TINY_DECODER = `${BASE_URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_decoder.pte`; +export const MOONSHINE_TINY_ENCODER = `${BASE_URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_encoder.pte`; +export const MOONSHINE_TOKENIZER = `${BASE_URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/moonshine_tiny_tokenizer.json`; +export const WHISPER_TOKENIZER = `${BASE_URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/whisper_tokenizer.json`; +export const WHISPER_TINY_DECODER = `${BASE_URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_decoder.pte`; +export const WHISPER_TINY_ENCODER = `${BASE_URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_encoder.pte`; +export const WHISPER_TINY_MULTILINGUAL_ENCODER = `${BASE_URL_PREFIX}-whisper-tiny/${VERSION_TAG}/xnnpack/xnnpack_whisper_encoder.pte`; +export const WHISPER_TINY_MULTILINGUAL_DECODER = `${BASE_URL_PREFIX}-whisper-tiny/${VERSION_TAG}/xnnpack/xnnpack_whisper_decoder.pte`; +export const WHISPER_TINY_MULTILINGUAL_TOKENIZER = `${BASE_URL_PREFIX}-whisper-tiny/${VERSION_TAG}/tokenizer.json`; // OCR -export const DETECTOR_CRAFT_1280 = - 'https://huggingface.co/software-mansion/react-native-executorch-detector-craft/resolve/v0.4.0/xnnpack/xnnpack_craft_1280.pte'; -export const DETECTOR_CRAFT_800 = - 'https://huggingface.co/software-mansion/react-native-executorch-detector-craft/resolve/v0.4.0/xnnpack/xnnpack_craft_800.pte'; -export const DETECTOR_CRAFT_320 = - 'https://huggingface.co/software-mansion/react-native-executorch-detector-craft/resolve/v0.4.0/xnnpack/xnnpack_craft_320.pte'; +export const DETECTOR_CRAFT_1280 = `${BASE_URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_1280.pte`; +export const DETECTOR_CRAFT_800 = `${BASE_URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_800.pte`; +export const DETECTOR_CRAFT_320 = `${BASE_URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_320.pte`; // Image segmentation -export const DEEPLAB_V3_RESNET50 = - 'https://huggingface.co/software-mansion/react-native-executorch-deeplab-v3/resolve/v0.4.0/xnnpack/deeplabV3_xnnpack_fp32.pte'; +export const DEEPLAB_V3_RESNET50 = `${BASE_URL_PREFIX}-deeplab-v3/${VERSION_TAG}/xnnpack/deeplabV3_xnnpack_fp32.pte`; // Image Embeddings export const CLIP_VIT_BASE_PATCH32_IMAGE_MODEL = 'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32/resolve/v0.5.0/clip-vit-base-patch32-vision_xnnpack.pte'; // Text Embeddings -export const ALL_MINILM_L6_V2 = - 'https://huggingface.co/software-mansion/react-native-executorch-all-MiniLM-L6-v2/resolve/v0.5.0/all-MiniLM-L6-v2_xnnpack.pte'; -export const ALL_MINILM_L6_V2_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-all-MiniLM-L6-v2/resolve/v0.5.0/tokenizer.json'; - -export const ALL_MPNET_BASE_V2 = - 'https://huggingface.co/software-mansion/react-native-executorch-all-mpnet-base-v2/resolve/v0.5.0/all-mpnet-base-v2_xnnpack.pte'; -export const ALL_MPNET_BASE_V2_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-all-mpnet-base-v2/resolve/v0.5.0/tokenizer.json'; - -export const MULTI_QA_MINILM_L6_COS_V1 = - 'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-MiniLM-L6-cos-v1/resolve/v0.5.0/multi-qa-MiniLM-L6-cos-v1_xnnpack.pte'; -export const MULTI_QA_MINILM_L6_COS_V1_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-MiniLM-L6-cos-v1/resolve/v0.5.0/tokenizer.json'; - -export const MULTI_QA_MPNET_BASE_DOT_V1 = - 'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-mpnet-base-dot-v1/resolve/v0.5.0/multi-qa-mpnet-base-dot-v1_xnnpack.pte'; -export const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-mpnet-base-dot-v1/resolve/v0.5.0/tokenizer.json'; +export const ALL_MINILM_L6_V2 = `${BASE_URL_PREFIX}-all-MiniLM-L6-v2/${VERSION_TAG}/all-MiniLM-L6-v2_xnnpack.pte`; +export const ALL_MINILM_L6_V2_TOKENIZER = `${BASE_URL_PREFIX}-all-MiniLM-L6-v2/${VERSION_TAG}/tokenizer.json`; + +export const ALL_MPNET_BASE_V2 = `${BASE_URL_PREFIX}-all-mpnet-base-v2/${VERSION_TAG}/all-mpnet-base-v2_xnnpack.pte`; +export const ALL_MPNET_BASE_V2_TOKENIZER = `${BASE_URL_PREFIX}-all-mpnet-base-v2/${VERSION_TAG}/tokenizer.json`; + +export const MULTI_QA_MINILM_L6_COS_V1 = `${BASE_URL_PREFIX}-multi-qa-MiniLM-L6-cos-v1/${VERSION_TAG}/multi-qa-MiniLM-L6-cos-v1_xnnpack.pte`; +export const MULTI_QA_MINILM_L6_COS_V1_TOKENIZER = `${BASE_URL_PREFIX}-multi-qa-MiniLM-L6-cos-v1/${VERSION_TAG}/tokenizer.json`; + +export const MULTI_QA_MPNET_BASE_DOT_V1 = `${BASE_URL_PREFIX}-multi-qa-mpnet-base-dot-v1/${VERSION_TAG}/multi-qa-mpnet-base-dot-v1_xnnpack.pte`; +export const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER = `${BASE_URL_PREFIX}-multi-qa-mpnet-base-dot-v1/${VERSION_TAG}/tokenizer.json`; export const CLIP_VIT_BASE_PATCH32_TEXT_MODEL = 'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32/resolve/v0.5.0/clip-vit-base-patch32-text_xnnpack.pte'; @@ -201,6 +139,7 @@ export const CLIP_VIT_BASE_PATCH32_IMAGE = { modelSource: CLIP_VIT_BASE_PATCH32_IMAGE_MODEL, }; + // Backward compatibility export const LLAMA3_2_3B_URL = LLAMA3_2_3B; export const LLAMA3_2_3B_QLORA_URL = LLAMA3_2_3B_QLORA; From af69af88c07e5f5600c26f78b722d42d55b5376e Mon Sep 17 00:00:00 2001 From: jakmro Date: Tue, 29 Jul 2025 17:10:24 +0200 Subject: [PATCH 02/20] refactor: change exported constants to object presets and update documentation to reflect these changes --- .../app/classification/index.tsx | 4 +- .../app/image_segmentation/index.tsx | 4 +- .../app/object_detection/index.tsx | 4 +- apps/computer-vision/app/ocr/index.tsx | 18 +- .../app/ocr_vertical/index.tsx | 18 +- .../app/style_transfer/index.tsx | 4 +- apps/llm/app/llm/index.tsx | 13 +- apps/llm/app/llm_structured_output/index.tsx | 13 +- apps/llm/app/llm_tool_calling/index.tsx | 10 +- apps/llm/app/voice_chat/index.tsx | 11 +- .../screens/SpeechToTextScreen.tsx | 4 +- .../app/text-embeddings/index.tsx | 11 +- .../01-natural-language-processing/useLLM.md | 45 +- .../useSpeechToText.md | 20 +- .../useTextEmbeddings.md | 22 +- .../useTokenizer.md | 18 +- .../02-computer-vision/useClassification.md | 8 +- .../useImageSegmentation.md | 8 +- .../02-hooks/02-computer-vision/useOCR.md | 36 +- .../02-computer-vision/useObjectDetection.md | 8 +- .../02-computer-vision/useStyleTransfer.md | 8 +- .../02-computer-vision/useVerticalOCR.md | 36 +- .../src/constants/modelUrls.ts | 452 ++++-- .../src/constants/ocr/models.ts | 1222 +++++++++++------ .../src/constants/ocr/symbols.ts | 126 +- .../src/constants/sttDefaults.ts | 30 +- .../src/controllers/SpeechToTextController.ts | 2 +- .../react-native-executorch/src/types/stt.ts | 6 +- 28 files changed, 1316 insertions(+), 845 deletions(-) diff --git a/apps/computer-vision/app/classification/index.tsx b/apps/computer-vision/app/classification/index.tsx index e473022540..19483d6996 100644 --- a/apps/computer-vision/app/classification/index.tsx +++ b/apps/computer-vision/app/classification/index.tsx @@ -13,9 +13,7 @@ export default function ClassificationScreen() { ); const [imageUri, setImageUri] = useState(''); - const model = useClassification({ - modelSource: EFFICIENTNET_V2_S, - }); + const model = useClassification(EFFICIENTNET_V2_S); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/computer-vision/app/image_segmentation/index.tsx b/apps/computer-vision/app/image_segmentation/index.tsx index 4b15f11074..29965d540a 100644 --- a/apps/computer-vision/app/image_segmentation/index.tsx +++ b/apps/computer-vision/app/image_segmentation/index.tsx @@ -61,9 +61,7 @@ const numberToColor: number[][] = [ ]; export default function ImageSegmentationScreen() { - const model = useImageSegmentation({ - modelSource: DEEPLAB_V3_RESNET50, - }); + const model = useImageSegmentation(DEEPLAB_V3_RESNET50); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/computer-vision/app/object_detection/index.tsx b/apps/computer-vision/app/object_detection/index.tsx index 4838418575..c0aed6a172 100644 --- a/apps/computer-vision/app/object_detection/index.tsx +++ b/apps/computer-vision/app/object_detection/index.tsx @@ -20,9 +20,7 @@ export default function ObjectDetectionScreen() { height: number; }>(); - const ssdLite = useObjectDetection({ - modelSource: SSDLITE_320_MOBILENET_V3_LARGE, - }); + const ssdLite = useObjectDetection(SSDLITE_320_MOBILENET_V3_LARGE); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(ssdLite.isGenerating); diff --git a/apps/computer-vision/app/ocr/index.tsx b/apps/computer-vision/app/ocr/index.tsx index 755f11b007..47dff1bbd5 100644 --- a/apps/computer-vision/app/ocr/index.tsx +++ b/apps/computer-vision/app/ocr/index.tsx @@ -1,13 +1,7 @@ import Spinner from 'react-native-loading-spinner-overlay'; import { BottomBar } from '../../components/BottomBar'; import { getImage } from '../../utils'; -import { - DETECTOR_CRAFT_800, - RECOGNIZER_EN_CRNN_128, - RECOGNIZER_EN_CRNN_256, - RECOGNIZER_EN_CRNN_512, - useOCR, -} from 'react-native-executorch'; +import { useOCR, OCR_ENGLISH } from 'react-native-executorch'; import { View, StyleSheet, Image, Text, ScrollView } from 'react-native'; import ImageWithBboxes2 from '../../components/ImageWithOCRBboxes'; import React, { useContext, useEffect, useState } from 'react'; @@ -22,15 +16,7 @@ export default function OCRScreen() { height: number; }>(); - const model = useOCR({ - detectorSource: DETECTOR_CRAFT_800, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerMedium: RECOGNIZER_EN_CRNN_256, - recognizerSmall: RECOGNIZER_EN_CRNN_128, - }, - language: 'en', - }); + const model = useOCR(OCR_ENGLISH); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/computer-vision/app/ocr_vertical/index.tsx b/apps/computer-vision/app/ocr_vertical/index.tsx index 333d431074..6879b81563 100644 --- a/apps/computer-vision/app/ocr_vertical/index.tsx +++ b/apps/computer-vision/app/ocr_vertical/index.tsx @@ -1,13 +1,7 @@ import Spinner from 'react-native-loading-spinner-overlay'; import { BottomBar } from '../../components/BottomBar'; import { getImage } from '../../utils'; -import { - DETECTOR_CRAFT_1280, - DETECTOR_CRAFT_320, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_64, - useVerticalOCR, -} from 'react-native-executorch'; +import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; import { View, StyleSheet, Image, Text, ScrollView } from 'react-native'; import ImageWithBboxes2 from '../../components/ImageWithOCRBboxes'; import React, { useContext, useEffect, useState } from 'react'; @@ -22,15 +16,7 @@ export default function VerticalOCRScree() { height: number; }>(); const model = useVerticalOCR({ - detectorSources: { - detectorLarge: DETECTOR_CRAFT_1280, - detectorNarrow: DETECTOR_CRAFT_320, - }, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerSmall: RECOGNIZER_EN_CRNN_64, - }, - language: 'en', + ...VERTICAL_OCR_ENGLISH, independentCharacters: true, }); const { setGlobalGenerating } = useContext(GeneratingContext); diff --git a/apps/computer-vision/app/style_transfer/index.tsx b/apps/computer-vision/app/style_transfer/index.tsx index 91bf71238c..7cec152ac7 100644 --- a/apps/computer-vision/app/style_transfer/index.tsx +++ b/apps/computer-vision/app/style_transfer/index.tsx @@ -11,9 +11,7 @@ import { GeneratingContext } from '../../context'; import ScreenWrapper from '../../ScreenWrapper'; export default function StyleTransferScreen() { - const model = useStyleTransfer({ - modelSource: STYLE_TRANSFER_CANDY, - }); + const model = useStyleTransfer(STYLE_TRANSFER_CANDY); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/llm/app/llm/index.tsx b/apps/llm/app/llm/index.tsx index f59e9e6894..b9eee28f2c 100644 --- a/apps/llm/app/llm/index.tsx +++ b/apps/llm/app/llm/index.tsx @@ -12,12 +12,7 @@ import { } from 'react-native'; import SendIcon from '../../assets/icons/send_icon.svg'; import Spinner from 'react-native-loading-spinner-overlay'; -import { - LLAMA3_2_1B_QLORA, - LLAMA3_2_TOKENIZER, - LLAMA3_2_TOKENIZER_CONFIG, - useLLM, -} from 'react-native-executorch'; +import { useLLM, LLAMA3_2_1B_QLORA } from 'react-native-executorch'; import PauseIcon from '../../assets/icons/pause_icon.svg'; import ColorPalette from '../../colors'; import Messages from '../../components/Messages'; @@ -35,11 +30,7 @@ function LLMScreen() { const textInputRef = useRef(null); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM({ - modelSource: LLAMA3_2_1B_QLORA, - tokenizerSource: LLAMA3_2_TOKENIZER, - tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, - }); + const llm = useLLM(LLAMA3_2_1B_QLORA); useEffect(() => { if (llm.error) { diff --git a/apps/llm/app/llm_structured_output/index.tsx b/apps/llm/app/llm_structured_output/index.tsx index dd6a7275be..9bdf6deb95 100644 --- a/apps/llm/app/llm_structured_output/index.tsx +++ b/apps/llm/app/llm_structured_output/index.tsx @@ -13,8 +13,6 @@ import { import SendIcon from '../../assets/icons/send_icon.svg'; import Spinner from 'react-native-loading-spinner-overlay'; import { - QWEN3_TOKENIZER, - QWEN3_TOKENIZER_CONFIG, useLLM, fixAndValidateStructuredOutput, getStructuredOutputPrompt, @@ -75,12 +73,7 @@ function LLMScreen() { const textInputRef = useRef(null); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM({ - // try out 4B model it this one struggles with following structured output - modelSource: QWEN3_1_7B_QUANTIZED, - tokenizerSource: QWEN3_TOKENIZER, - tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, - }); + const llm = useLLM(QWEN3_1_7B_QUANTIZED); // try out 4B model if 1.7B struggles with following structured output useEffect(() => { setGlobalGenerating(llm.isGenerating); @@ -89,10 +82,6 @@ function LLMScreen() { const { configure } = llm; useEffect(() => { const formattingInstructions = getStructuredOutputPrompt(responseSchema); - // const formattingInstructionsWithZod = getStructuredOutputPrompt( - // responseSchemaWithZod - // ); - const prompt = `Your goal is to parse user's messages and return them in JSON format. Don't respond to user. Simply return JSON with user's question parsed. \n${formattingInstructions}\n /no_think`; configure({ diff --git a/apps/llm/app/llm_tool_calling/index.tsx b/apps/llm/app/llm_tool_calling/index.tsx index 5d5995f882..5e1440ceab 100644 --- a/apps/llm/app/llm_tool_calling/index.tsx +++ b/apps/llm/app/llm_tool_calling/index.tsx @@ -14,11 +14,9 @@ import SWMIcon from '../../assets/icons/swm_icon.svg'; import SendIcon from '../../assets/icons/send_icon.svg'; import Spinner from 'react-native-loading-spinner-overlay'; import { - HAMMER2_1_1_5B, - HAMMER2_1_TOKENIZER, - HAMMER2_1_TOKENIZER_CONFIG, useLLM, DEFAULT_SYSTEM_PROMPT, + HAMMER2_1_1_5B_QUANTIZED, } from 'react-native-executorch'; import PauseIcon from '../../assets/icons/pause_icon.svg'; import ColorPalette from '../../colors'; @@ -41,11 +39,7 @@ function LLMToolCallingScreen() { const textInputRef = useRef(null); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM({ - modelSource: HAMMER2_1_1_5B, - tokenizerSource: HAMMER2_1_TOKENIZER, - tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, - }); + const llm = useLLM(HAMMER2_1_1_5B_QUANTIZED); useEffect(() => { setGlobalGenerating(llm.isGenerating); diff --git a/apps/llm/app/voice_chat/index.tsx b/apps/llm/app/voice_chat/index.tsx index c4baaa9870..67a7294e00 100644 --- a/apps/llm/app/voice_chat/index.tsx +++ b/apps/llm/app/voice_chat/index.tsx @@ -16,8 +16,7 @@ import { useSpeechToText, useLLM, QWEN3_0_6B_QUANTIZED, - QWEN3_TOKENIZER, - QWEN3_TOKENIZER_CONFIG, + MOONSHINE_TINY, } from 'react-native-executorch'; import PauseIcon from '../../assets/icons/pause_icon.svg'; import MicIcon from '../../assets/icons/mic_icon.svg'; @@ -68,13 +67,9 @@ function VoiceChatScreen() { const messageRecorded = useRef(false); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM({ - modelSource: QWEN3_0_6B_QUANTIZED, - tokenizerSource: QWEN3_TOKENIZER, - tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, - }); + const llm = useLLM(QWEN3_0_6B_QUANTIZED); const speechToText = useSpeechToText({ - modelName: 'moonshine', + ...MOONSHINE_TINY, windowSize: 3, overlapSeconds: 1.2, }); diff --git a/apps/speech-to-text/screens/SpeechToTextScreen.tsx b/apps/speech-to-text/screens/SpeechToTextScreen.tsx index 188c8cc50c..bae8f0998a 100644 --- a/apps/speech-to-text/screens/SpeechToTextScreen.tsx +++ b/apps/speech-to-text/screens/SpeechToTextScreen.tsx @@ -1,4 +1,4 @@ -import { useSpeechToText } from 'react-native-executorch'; +import { MOONSHINE_TINY, useSpeechToText } from 'react-native-executorch'; import React from 'react'; import { Text, @@ -54,7 +54,7 @@ export const SpeechToTextScreen = () => { sequence, error, transcribe, - } = useSpeechToText({ modelName: 'moonshine', streamingConfig: 'balanced' }); + } = useSpeechToText(MOONSHINE_TINY); const loadAudio = async (url: string) => { const audioContext = new AudioContext({ sampleRate: 16e3 }); diff --git a/apps/text-embeddings/app/text-embeddings/index.tsx b/apps/text-embeddings/app/text-embeddings/index.tsx index bc742f1dfa..9c6d9b6b93 100644 --- a/apps/text-embeddings/app/text-embeddings/index.tsx +++ b/apps/text-embeddings/app/text-embeddings/index.tsx @@ -11,11 +11,7 @@ import { Platform, } from 'react-native'; import { Ionicons } from '@expo/vector-icons'; -import { - useTextEmbeddings, - ALL_MINILM_L6_V2, - ALL_MINILM_L6_V2_TOKENIZER, -} from 'react-native-executorch'; +import { useTextEmbeddings, ALL_MINILM_L6_V2 } from 'react-native-executorch'; import { useIsFocused } from '@react-navigation/native'; import { dotProduct } from '../../utils/math'; @@ -26,10 +22,7 @@ export default function TextEmbeddingsScreenWrapper() { } function TextEmbeddingsScreen() { - const model = useTextEmbeddings({ - modelSource: ALL_MINILM_L6_V2, - tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER, - }); + const model = useTextEmbeddings(ALL_MINILM_L6_V2); const [inputSentence, setInputSentence] = useState(''); const [sentencesWithEmbeddings, setSentencesWithEmbeddings] = useState< diff --git a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md index a24dcf04ea..927b31aba8 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md @@ -39,18 +39,9 @@ Given computational constraints, our architecture is designed to support only on In order to load a model into the app, you need to run the following code: ```typescript -import { - useLLM, - LLAMA3_2_1B, - LLAMA3_2_TOKENIZER, - LLAMA3_2_TOKENIZER_CONFIG, -} from 'react-native-executorch'; - -const llm = useLLM({ - modelSource: LLAMA3_2_1B, - tokenizerSource: LLAMA3_2_TOKENIZER, - tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, -}); +import { useLLM, LLAMA3_2_1B } from 'react-native-executorch'; + +const llm = useLLM(LLAMA3_2_1B); ```
@@ -167,11 +158,7 @@ You can use functions returned from this hooks in two manners: To perform chat completion you can use the `generate` function. There is no return value. Instead, the `response` value is updated with each token. ```tsx -const llm = useLLM({ - modelSource: LLAMA3_2_1B, - tokenizerSource: LLAMA3_2_TOKENIZER, - tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, -}); +const llm = useLLM(LLAMA3_2_1B); const handleGenerate = () => { const chat = [ @@ -226,11 +213,7 @@ const TOOL_DEFINITIONS: LLMTool[] = [ }, ]; -const llm = useLLM({ - modelSource: HAMMER2_1_1_5B, - tokenizerSource: HAMMER2_1_1_5B_TOKENIZER, - tokenizerConfigSource: HAMMER2_1_1_5B_TOKENIZER_CONFIG, -}); +const llm = useLLM(HAMMER2_1_1_5B); const handleGenerate = () => { const chat = [ @@ -289,11 +272,7 @@ To configure model (i.e. change system prompt, load initial conversation history In order to send a message to the model, one can use the following code: ```tsx -const llm = useLLM({ - modelSource: LLAMA3_2_1B, - tokenizerSource: LLAMA3_2_TOKENIZER, - tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, -}); +const llm = useLLM(LLAMA3_2_1B); const send = () => { const message = 'Hi, who are you?'; @@ -338,11 +317,7 @@ const TOOL_DEFINITIONS: LLMTool[] = [ }, ]; -const llm = useLLM({ - modelSource: HAMMER2_1_1_5B, - tokenizerSource: HAMMER2_1_1_5B_TOKENIZER, - tokenizerConfigSource: HAMMER2_1_1_5B_TOKENIZER_CONFIG, -}); +const llm = useLLM(HAMMER2_1_1_5B); useEffect(() => { llm.configure({ @@ -418,11 +393,7 @@ const responseSchemaWithZod = z.object({ currency: z.optional(z.string().meta({ description: 'Currency of offer.' })), }); -const llm = useLLM({ - modelSource: QWEN3_4B_QUANTIZED, - tokenizerSource: QWEN3_TOKENIZER, - tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, -}); +const llm = useLLM(QWEN3_4B_QUANTIZED); useEffect(() => { const formattingInstructions = getStructuredOutputPrompt(responseSchema); diff --git a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md index 4115408fe2..710399c32a 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md @@ -33,13 +33,11 @@ It is recommended to use models provided by us, which are available at our [Hugg You can obtain waveform from audio in any way most suitable to you, however in the snippet below we utilize `react-native-audio-api` library to process a mp3 file. ```typescript -import { useSpeechToText } from 'react-native-executorch'; +import { useSpeechToText, MOONSHINE_TINY } from 'react-native-executorch'; import { AudioContext } from 'react-native-audio-api'; import * as FileSystem from 'expo-file-system'; -const { transcribe, error } = useSpeechToText({ - modelName: 'moonshine', -}); +const { transcribe, error } = useSpeechToText(MOONSHINE_TINY); const loadAudio = async (url: string) => { const audioContext = new AudioContext({ sampleRate: 16e3 }); @@ -216,14 +214,12 @@ await model.transcribe(mySpanishAudio, SpeechToTextLanguage.Spanish); ```tsx import { Button, Text, View } from 'react-native'; -import { useSpeechToText } from 'react-native-executorch'; +import { useSpeechToText, WHISPER_TINY } from 'react-native-executorch'; import * as FileSystem from 'expo-file-system'; import { AudioContext } from 'react-native-audio-api'; function App() { - const { transcribe, sequence, error } = useSpeechToText({ - modelName: 'whisper', - }); + const { transcribe, sequence, error } = useSpeechToText(WHISPER_TINY); const loadAudio = async (url: string) => { const audioContext = new AudioContext({ sampleRate: 16e3 }); @@ -255,7 +251,11 @@ function App() { ### Live data (microphone) transcription ```tsx -import { STREAMING_ACTION, useSpeechToText } from 'react-native-executorch'; +import { + STREAMING_ACTION, + useSpeechToText, + MOONSHINE_TINY, +} from 'react-native-executorch'; import LiveAudioStream from 'react-native-live-audio-stream'; import { useState } from 'react'; import { Buffer } from 'buffer'; @@ -291,7 +291,7 @@ const float32ArrayFromPCMBinaryBuffer = (b64EncodedBuffer: string) => { function App() { const [isRecording, setIsRecording] = useState(false); const speechToText = useSpeechToText({ - modelName: 'moonshine', + ...MOONSHINE_TINY, windowSize: 3, overlapSeconds: 1.2, }); diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md index a8dfcbdc2d..805ad9341f 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md @@ -24,16 +24,9 @@ It is recommended to use models provided by us, which are available at our [Hugg ## Reference ```typescript -import { - useTextEmbeddings, - ALL_MINILM_L6_V2, - ALL_MINILM_L6_V2_TOKENIZER, -} from 'react-native-executorch'; +import { useTextEmbeddings, ALL_MINILM_L6_V2 } from 'react-native-executorch'; -const model = useTextEmbeddings({ - modelSource: ALL_MINILM_L6_V2, - tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER, -}); +const model = useTextEmbeddings(ALL_MINILM_L6_V2); try { const embedding = await model.forward('Hello World!'); @@ -69,11 +62,7 @@ To run the model, you can use the `forward` method. It accepts one argument, whi ## Example ```typescript -import { - useTextEmbeddings, - ALL_MINILM_L6_V2, - ALL_MINILM_L6_V2_TOKENIZER, -} from 'react-native-executorch'; +import { useTextEmbeddings, ALL_MINILM_L6_V2 } from 'react-native-executorch'; const dotProduct = (a: number[], b: number[]) => a.reduce((sum, val, i) => sum + val * b[i], 0); @@ -86,10 +75,7 @@ const cosineSimilarity = (a: number[], b: number[]) => { }; function App() { - const model = useTextEmbeddings({ - modelSource: ALL_MINILM_L6_V2, - tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER, - }); + const model = useTextEmbeddings(ALL_MINILM_L6_V2); // ... diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md index a5c6cd7549..74fb8e54ea 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md @@ -25,14 +25,9 @@ We are using [Hugging Face Tokenizers](https://huggingface.co/docs/tokenizers/in ## Reference ```typescript -import { - useTokenizer, - ALL_MINILM_L6_V2_TOKENIZER, -} from 'react-native-executorch'; +import { useTokenizer, ALL_MINILM_L6_V2 } from 'react-native-executorch'; -const tokenizer = useTokenizer({ - tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER, -}); +const tokenizer = useTokenizer(ALL_MINILM_L6_V2); const text = 'Hello, world!'; @@ -72,15 +67,10 @@ try { ## Example ```typescript -import { - useTokenizer, - ALL_MINILM_L6_V2_TOKENIZER, -} from 'react-native-executorch'; +import { useTokenizer, ALL_MINILM_L6_V2 } from 'react-native-executorch'; function App() { - const tokenizer = useTokenizer({ - tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER, - }); + const tokenizer = useTokenizer(ALL_MINILM_L6_V2); // ... diff --git a/docs/docs/02-hooks/02-computer-vision/useClassification.md b/docs/docs/02-hooks/02-computer-vision/useClassification.md index 30f8c6390a..be702e3e6f 100644 --- a/docs/docs/02-hooks/02-computer-vision/useClassification.md +++ b/docs/docs/02-hooks/02-computer-vision/useClassification.md @@ -17,9 +17,7 @@ It is recommended to use models provided by us, which are available at our [Hugg ```typescript import { useClassification, EFFICIENTNET_V2_S } from 'react-native-executorch'; -const model = useClassification({ - modelSource: EFFICIENTNET_V2_S, -}); +const model = useClassification(EFFICIENTNET_V2_S); const imageUri = 'file::///Users/.../cute_puppy.png'; @@ -61,9 +59,7 @@ Images from external sources are stored in your application's temporary director import { useClassification, EFFICIENTNET_V2_S } from 'react-native-executorch'; function App() { - const model = useClassification({ - modulePath: EFFICIENTNET_V2_S, - }); + const model = useClassification(EFFICIENTNET_V2_S); // ... const imageUri = 'file:///Users/.../cute_puppy.png'; diff --git a/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md b/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md index 8f43c87d69..e2b1c1a11b 100644 --- a/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md +++ b/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md @@ -16,9 +16,7 @@ import { DEEPLAB_V3_RESNET50, } from 'react-native-executorch'; -const model = useImageSegmentation({ - modelSource: DEEPLAB_V3_RESNET50, -}); +const model = useImageSegmentation(DEEPLAB_V3_RESNET50); const imageUri = 'file::///Users/.../cute_cat.png'; @@ -67,9 +65,7 @@ Setting `resize` to true will make `forward` slower. ```typescript function App() { - const model = useImageSegmentation({ - modelSource: DEEPLAB_V3_RESNET50, - }); + const model = useImageSegmentation(DEEPLAB_V3_RESNET50); // ... const imageUri = 'file::///Users/.../cute_cat.png'; diff --git a/docs/docs/02-hooks/02-computer-vision/useOCR.md b/docs/docs/02-hooks/02-computer-vision/useOCR.md index 93017354fb..1e757b281f 100644 --- a/docs/docs/02-hooks/02-computer-vision/useOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useOCR.md @@ -11,24 +11,10 @@ It is recommended to use models provided by us, which are available at our [Hugg ## Reference ```tsx -import { - useOCR, - DETECTOR_CRAFT_800, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_256, - RECOGNIZER_EN_CRNN_128, -} from 'react-native-executorch'; +import { useOCR, OCR_ENGLISH } from 'react-native-executorch'; function App() { - const model = useOCR({ - detectorSource: DETECTOR_CRAFT_800, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerMedium: RECOGNIZER_EN_CRNN_256, - recognizerSmall: RECOGNIZER_EN_CRNN_128, - }, - language: 'en', - }); + const model = useOCR(OCR_ENGLISH); // ... for (const ocrDetection of await model.forward('https://url-to-image.jpg')) { @@ -185,24 +171,10 @@ The `text` property contains the text recognized within detected text region. Th ## Example ```tsx -import { - useOCR, - DETECTOR_CRAFT_800, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_256, - RECOGNIZER_EN_CRNN_128, -} from 'react-native-executorch'; +import { useOCR, OCR_ENGLISH } from 'react-native-executorch'; function App() { - const model = useOCR({ - detectorSource: DETECTOR_CRAFT_800, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerMedium: RECOGNIZER_EN_CRNN_256, - recognizerSmall: RECOGNIZER_EN_CRNN_128, - }, - language: 'en', - }); + const model = useOCR(OCR_ENGLISH); const runModel = async () => { const ocrDetections = await model.forward('https://url-to-image.jpg'); diff --git a/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md b/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md index 53a48369b2..6f615a20c7 100644 --- a/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md +++ b/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md @@ -18,9 +18,7 @@ import { } from 'react-native-executorch'; function App() { - const ssdlite = useObjectDetection({ - modelSource: SSDLITE_320_MOBILENET_V3_LARGE, // alternatively, you can use require(...) - }); + const ssdlite = useObjectDetection(SSDLITE_320_MOBILENET_V3_LARGE); // ... for (const detection of await ssdlite.forward('https://url-to-image.jpg')) { @@ -106,9 +104,7 @@ import { } from 'react-native-executorch'; function App() { - const ssdlite = useObjectDetection({ - modelSource: SSDLITE_320_MOBILENET_V3_LARGE, - }); + const ssdlite = useObjectDetection(SSDLITE_320_MOBILENET_V3_LARGE); const runModel = async () => { const detections = await ssdlite.forward('https://url-to-image.jpg'); diff --git a/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md b/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md index 3fefe6d934..44d4c31ba2 100644 --- a/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md +++ b/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md @@ -16,9 +16,7 @@ import { STYLE_TRANSFER_CANDY, } from 'react-native-executorch'; -const model = useStyleTransfer({ - modelSource: STYLE_TRANSFER_CANDY, -}); +const model = useStyleTransfer(STYLE_TRANSFER_CANDY); const imageUri = 'file::///Users/.../cute_cat.png'; @@ -58,9 +56,7 @@ Images from external sources and the generated image are stored in your applicat ```typescript function App() { - const model = useStyleTransfer({ - modelSource: STYLE_TRANSFER_CANDY, - }); + const model = useStyleTransfer(STYLE_TRANSFER_CANDY); // ... const imageUri = 'file::///Users/.../cute_cat.png'; diff --git a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md index 1cd8668dcc..ea285bc32a 100644 --- a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md @@ -15,25 +15,11 @@ It is recommended to use models provided by us, which are available at our [Hugg ## Reference ```tsx -import { - DETECTOR_CRAFT_1280, - DETECTOR_CRAFT_320, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_64, - useVerticalOCR, -} from 'react-native-executorch'; +import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; function App() { const model = useVerticalOCR({ - detectorSources: { - detectorLarge: DETECTOR_CRAFT_1280, - detectorNarrow: DETECTOR_CRAFT_320, - }, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerSmall: RECOGNIZER_EN_CRNN_64, - }, - language: 'en', + ...VERTICAL_OCR_ENGLISH, independentCharacters: true, }); @@ -202,25 +188,11 @@ The `text` property contains the text recognized within detected text region. Th ## Example ```tsx -import { - DETECTOR_CRAFT_1280, - DETECTOR_CRAFT_320, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_64, - useVerticalOCR, -} from 'react-native-executorch'; +import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; function App() { const model = useVerticalOCR({ - detectorSources: { - detectorLarge: DETECTOR_CRAFT_1280, - detectorNarrow: DETECTOR_CRAFT_320, - }, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerSmall: RECOGNIZER_EN_CRNN_64, - }, - language: 'en', + ...VERTICAL_OCR_ENGLISH, independentCharacters: true, }); diff --git a/packages/react-native-executorch/src/constants/modelUrls.ts b/packages/react-native-executorch/src/constants/modelUrls.ts index e68ebebbce..574e1945be 100644 --- a/packages/react-native-executorch/src/constants/modelUrls.ts +++ b/packages/react-native-executorch/src/constants/modelUrls.ts @@ -1,151 +1,391 @@ import { Platform } from 'react-native'; +import { AvailableModels } from '../types/stt'; -const BASE_URL_PREFIX = +const URL_PREFIX = 'https://huggingface.co/software-mansion/react-native-executorch'; const VERSION_TAG = 'resolve/v0.4.0'; +const NEXT_VERSION_TAG = 'resolve/v0.5.0'; -// LLM`s +// LLMs // LLAMA 3.2 -export const LLAMA3_2_3B = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/original/llama3_2_3B_bf16.pte`; -export const LLAMA3_2_3B_QLORA = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/QLoRA/llama3_2-3B_qat_lora.pte`; -export const LLAMA3_2_3B_SPINQUANT = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/spinquant/llama3_2_3B_spinquant.pte`; -export const LLAMA3_2_1B = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/original/llama3_2_bf16.pte`; -export const LLAMA3_2_1B_QLORA = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/QLoRA/llama3_2_qat_lora.pte`; -export const LLAMA3_2_1B_SPINQUANT = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/spinquant/llama3_2_spinquant.pte`; -export const LLAMA3_2_TOKENIZER = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/tokenizer.json`; -export const LLAMA3_2_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-llama-3.2/${VERSION_TAG}/tokenizer_config.json`; +const LLAMA3_2_3B_MODEL = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/original/llama3_2_3B_bf16.pte`; +const LLAMA3_2_3B_QLORA_MODEL = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/QLoRA/llama3_2-3B_qat_lora.pte`; +const LLAMA3_2_3B_SPINQUANT_MODEL = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-3B/spinquant/llama3_2_3B_spinquant.pte`; +const LLAMA3_2_1B_MODEL = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/original/llama3_2_bf16.pte`; +const LLAMA3_2_1B_QLORA_MODEL = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/QLoRA/llama3_2_qat_lora.pte`; +const LLAMA3_2_1B_SPINQUANT_MODEL = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/llama-3.2-1B/spinquant/llama3_2_spinquant.pte`; +const LLAMA3_2_TOKENIZER = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/tokenizer.json`; +const LLAMA3_2_TOKENIZER_CONFIG = `${URL_PREFIX}-llama-3.2/${VERSION_TAG}/tokenizer_config.json`; + +export const LLAMA3_2_3B = { + modelSource: LLAMA3_2_3B_MODEL, + tokenizerSource: LLAMA3_2_TOKENIZER, + tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, +}; + +export const LLAMA3_2_3B_QLORA = { + modelSource: LLAMA3_2_3B_QLORA_MODEL, + tokenizerSource: LLAMA3_2_TOKENIZER, + tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, +}; + +export const LLAMA3_2_3B_SPINQUANT = { + modelSource: LLAMA3_2_3B_SPINQUANT_MODEL, + tokenizerSource: LLAMA3_2_TOKENIZER, + tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, +}; + +export const LLAMA3_2_1B = { + modelSource: LLAMA3_2_1B_MODEL, + tokenizerSource: LLAMA3_2_TOKENIZER, + tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, +}; + +export const LLAMA3_2_1B_QLORA = { + modelSource: LLAMA3_2_1B_QLORA_MODEL, + tokenizerSource: LLAMA3_2_TOKENIZER, + tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, +}; + +export const LLAMA3_2_1B_SPINQUANT = { + modelSource: LLAMA3_2_1B_SPINQUANT_MODEL, + tokenizerSource: LLAMA3_2_TOKENIZER, + tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, +}; // QWEN 3 -export const QWEN3_0_6B = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-0.6B/original/qwen3_0_6b_bf16.pte`; -export const QWEN3_0_6B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-0.6B/quantized/qwen3_0_6b_8da4w.pte`; -export const QWEN3_1_7B = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-1.7B/original/qwen3_1_7b_bf16.pte`; -export const QWEN3_1_7B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-1.7B/quantized/qwen3_1_7b_8da4w.pte`; -export const QWEN3_4B = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-4B/original/qwen3_4b_bf16.pte`; -export const QWEN3_4B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-4B/quantized/qwen3_4b_8da4w.pte`; -export const QWEN3_TOKENIZER = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/tokenizer.json`; -export const QWEN3_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-qwen-3/${VERSION_TAG}/tokenizer_config.json`; +const QWEN3_0_6B_MODEL = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-0.6B/original/qwen3_0_6b_bf16.pte`; +const QWEN3_0_6B_QUANTIZED_MODEL = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-0.6B/quantized/qwen3_0_6b_8da4w.pte`; +const QWEN3_1_7B_MODEL = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-1.7B/original/qwen3_1_7b_bf16.pte`; +const QWEN3_1_7B_QUANTIZED_MODEL = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-1.7B/quantized/qwen3_1_7b_8da4w.pte`; +const QWEN3_4B_MODEL = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-4B/original/qwen3_4b_bf16.pte`; +const QWEN3_4B_QUANTIZED_MODEL = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/qwen-3-4B/quantized/qwen3_4b_8da4w.pte`; +const QWEN3_TOKENIZER = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/tokenizer.json`; +const QWEN3_TOKENIZER_CONFIG = `${URL_PREFIX}-qwen-3/${VERSION_TAG}/tokenizer_config.json`; + +export const QWEN3_0_6B = { + modelSource: QWEN3_0_6B_MODEL, + tokenizerSource: QWEN3_TOKENIZER, + tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, +}; + +export const QWEN3_0_6B_QUANTIZED = { + modelSource: QWEN3_0_6B_QUANTIZED_MODEL, + tokenizerSource: QWEN3_TOKENIZER, + tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, +}; + +export const QWEN3_1_7B = { + modelSource: QWEN3_1_7B_MODEL, + tokenizerSource: QWEN3_TOKENIZER, + tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, +}; + +export const QWEN3_1_7B_QUANTIZED = { + modelSource: QWEN3_1_7B_QUANTIZED_MODEL, + tokenizerSource: QWEN3_TOKENIZER, + tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, +}; + +export const QWEN3_4B = { + modelSource: QWEN3_4B_MODEL, + tokenizerSource: QWEN3_TOKENIZER, + tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, +}; + +export const QWEN3_4B_QUANTIZED = { + modelSource: QWEN3_4B_QUANTIZED_MODEL, + tokenizerSource: QWEN3_TOKENIZER, + tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG, +}; // HAMMER 2.1 -export const HAMMER2_1_0_5B = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-0.5B/original/hammer2_1_0_5B_bf16.pte`; -export const HAMMER2_1_0_5B_QUANTIZED = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-0.5B/quantized/hammer2_1_0_5B_8da4w.pte`; -export const HAMMER2_1_1_5B = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-1.5B/original/hammer2_1_1_5B_bf16.pte`; -export const HAMMER2_1_1_5B_QUANTIZED = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-1.5B/quantized/hammer2_1_1_5B_8da4w.pte`; -export const HAMMER2_1_3B = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-3B/original/hammer2_1_3B_bf16.pte`; -export const HAMMER2_1_3B_QUANTIZED = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-3B/quantized/hammer2_1_3B_8da4w.pte`; -export const HAMMER2_1_TOKENIZER = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/tokenizer.json`; -export const HAMMER2_1_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-hammer-2.1/${VERSION_TAG}/tokenizer_config.json`; +const HAMMER2_1_0_5B_MODEL = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-0.5B/original/hammer2_1_0_5B_bf16.pte`; +const HAMMER2_1_0_5B_QUANTIZED_MODEL = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-0.5B/quantized/hammer2_1_0_5B_8da4w.pte`; +const HAMMER2_1_1_5B_MODEL = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-1.5B/original/hammer2_1_1_5B_bf16.pte`; +const HAMMER2_1_1_5B_QUANTIZED_MODEL = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-1.5B/quantized/hammer2_1_1_5B_8da4w.pte`; +const HAMMER2_1_3B_MODEL = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-3B/original/hammer2_1_3B_bf16.pte`; +const HAMMER2_1_3B_QUANTIZED_MODEL = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/hammer-2.1-3B/quantized/hammer2_1_3B_8da4w.pte`; +const HAMMER2_1_TOKENIZER = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/tokenizer.json`; +const HAMMER2_1_TOKENIZER_CONFIG = `${URL_PREFIX}-hammer-2.1/${VERSION_TAG}/tokenizer_config.json`; + +export const HAMMER2_1_0_5B = { + modelSource: HAMMER2_1_0_5B_MODEL, + tokenizerSource: HAMMER2_1_TOKENIZER, + tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, +}; + +export const HAMMER2_1_0_5B_QUANTIZED = { + modelSource: HAMMER2_1_0_5B_QUANTIZED_MODEL, + tokenizerSource: HAMMER2_1_TOKENIZER, + tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, +}; + +export const HAMMER2_1_1_5B = { + modelSource: HAMMER2_1_1_5B_MODEL, + tokenizerSource: HAMMER2_1_TOKENIZER, + tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, +}; + +export const HAMMER2_1_1_5B_QUANTIZED = { + modelSource: HAMMER2_1_1_5B_QUANTIZED_MODEL, + tokenizerSource: HAMMER2_1_TOKENIZER, + tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, +}; + +export const HAMMER2_1_3B = { + modelSource: HAMMER2_1_3B_MODEL, + tokenizerSource: HAMMER2_1_TOKENIZER, + tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, +}; + +export const HAMMER2_1_3B_QUANTIZED = { + modelSource: HAMMER2_1_3B_QUANTIZED_MODEL, + tokenizerSource: HAMMER2_1_TOKENIZER, + tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG, +}; // SMOLLM2 -export const SMOLLM2_1_135M = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-135M/original/smolLm2_135M_bf16.pte`; -export const SMOLLM2_1_135M_QUANTIZED = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-135M/quantized/smolLm2_135M_8da4w.pte`; -export const SMOLLM2_1_360M = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-360M/original/smolLm2_360M_bf16.pte`; -export const SMOLLM2_1_360M_QUANTIZED = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-360M/quantized/smolLm2_360M_8da4w.pte`; -export const SMOLLM2_1_1_7B = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-1.7B/original/smolLm2_1_7B_bf16.pte`; -export const SMOLLM2_1_1_7B_QUANTIZED = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-1.7B/quantized/smolLm2_1_7B_8da4w.pte`; -export const SMOLLM2_1_TOKENIZER = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/tokenizer.json`; -export const SMOLLM2_1_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-smolLm-2/${VERSION_TAG}/tokenizer_config.json`; +const SMOLLM2_1_135M_MODEL = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-135M/original/smolLm2_135M_bf16.pte`; +const SMOLLM2_1_135M_QUANTIZED_MODEL = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-135M/quantized/smolLm2_135M_8da4w.pte`; +const SMOLLM2_1_360M_MODEL = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-360M/original/smolLm2_360M_bf16.pte`; +const SMOLLM2_1_360M_QUANTIZED_MODEL = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-360M/quantized/smolLm2_360M_8da4w.pte`; +const SMOLLM2_1_1_7B_MODEL = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-1.7B/original/smolLm2_1_7B_bf16.pte`; +const SMOLLM2_1_1_7B_QUANTIZED_MODEL = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/smolLm-2-1.7B/quantized/smolLm2_1_7B_8da4w.pte`; +const SMOLLM2_1_TOKENIZER = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/tokenizer.json`; +const SMOLLM2_1_TOKENIZER_CONFIG = `${URL_PREFIX}-smolLm-2/${VERSION_TAG}/tokenizer_config.json`; + +export const SMOLLM2_1_135M = { + modelSource: SMOLLM2_1_135M_MODEL, + tokenizerSource: SMOLLM2_1_TOKENIZER, + tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG, +}; + +export const SMOLLM2_1_135M_QUANTIZED = { + modelSource: SMOLLM2_1_135M_QUANTIZED_MODEL, + tokenizerSource: SMOLLM2_1_TOKENIZER, + tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG, +}; + +export const SMOLLM2_1_360M = { + modelSource: SMOLLM2_1_360M_MODEL, + tokenizerSource: SMOLLM2_1_TOKENIZER, + tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG, +}; + +export const SMOLLM2_1_360M_QUANTIZED = { + modelSource: SMOLLM2_1_360M_QUANTIZED_MODEL, + tokenizerSource: SMOLLM2_1_TOKENIZER, + tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG, +}; + +export const SMOLLM2_1_1_7B = { + modelSource: SMOLLM2_1_1_7B_MODEL, + tokenizerSource: SMOLLM2_1_TOKENIZER, + tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG, +}; + +export const SMOLLM2_1_1_7B_QUANTIZED = { + modelSource: SMOLLM2_1_1_7B_QUANTIZED_MODEL, + tokenizerSource: SMOLLM2_1_TOKENIZER, + tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG, +}; // QWEN 2.5 -export const QWEN2_5_0_5B = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-0.5B/original/qwen2_5_0_5b_bf16.pte`; -export const QWEN2_5_0_5B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-0.5B/quantized/qwen2_5_0_5b_8da4w.pte`; -export const QWEN2_5_1_5B = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-1.5B/original/qwen2_5_1_5b_bf16.pte`; -export const QWEN2_5_1_5B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-1.5B/quantized/qwen2_5_1_5b_8da4w.pte`; -export const QWEN2_5_3B = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-3B/original/qwen2_5_3b_bf16.pte`; -export const QWEN2_5_3B_QUANTIZED = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-3B/quantized/qwen2_5_3b_8da4w.pte`; -export const QWEN2_5_TOKENIZER = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/tokenizer.json`; -export const QWEN2_5_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-qwen-2.5/${VERSION_TAG}/tokenizer_config.json`; +const QWEN2_5_0_5B_MODEL = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-0.5B/original/qwen2_5_0_5b_bf16.pte`; +const QWEN2_5_0_5B_QUANTIZED_MODEL = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-0.5B/quantized/qwen2_5_0_5b_8da4w.pte`; +const QWEN2_5_1_5B_MODEL = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-1.5B/original/qwen2_5_1_5b_bf16.pte`; +const QWEN2_5_1_5B_QUANTIZED_MODEL = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-1.5B/quantized/qwen2_5_1_5b_8da4w.pte`; +const QWEN2_5_3B_MODEL = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-3B/original/qwen2_5_3b_bf16.pte`; +const QWEN2_5_3B_QUANTIZED_MODEL = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/qwen-2.5-3B/quantized/qwen2_5_3b_8da4w.pte`; +const QWEN2_5_TOKENIZER = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/tokenizer.json`; +const QWEN2_5_TOKENIZER_CONFIG = `${URL_PREFIX}-qwen-2.5/${VERSION_TAG}/tokenizer_config.json`; + +export const QWEN2_5_0_5B = { + modelSource: QWEN2_5_0_5B_MODEL, + tokenizerSource: QWEN2_5_TOKENIZER, + tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG, +}; + +export const QWEN2_5_0_5B_QUANTIZED = { + modelSource: QWEN2_5_0_5B_QUANTIZED_MODEL, + tokenizerSource: QWEN2_5_TOKENIZER, + tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG, +}; + +export const QWEN2_5_1_5B = { + modelSource: QWEN2_5_1_5B_MODEL, + tokenizerSource: QWEN2_5_TOKENIZER, + tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG, +}; + +export const QWEN2_5_1_5B_QUANTIZED = { + modelSource: QWEN2_5_1_5B_QUANTIZED_MODEL, + tokenizerSource: QWEN2_5_TOKENIZER, + tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG, +}; + +export const QWEN2_5_3B = { + modelSource: QWEN2_5_3B_MODEL, + tokenizerSource: QWEN2_5_TOKENIZER, + tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG, +}; + +export const QWEN2_5_3B_QUANTIZED = { + modelSource: QWEN2_5_3B_QUANTIZED_MODEL, + tokenizerSource: QWEN2_5_TOKENIZER, + tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG, +}; // PHI 4 -export const PHI_4_MINI_4B = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/original/phi-4-mini_bf16.pte`; -export const PHI_4_MINI_4B_QUANTIZED = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/quantized/phi-4-mini_8da4w.pte`; -export const PHI_4_MINI_TOKENIZER = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/tokenizer.json`; -export const PHI_4_MINI_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-phi-4-mini/${VERSION_TAG}/tokenizer_config.json`; +const PHI_4_MINI_4B_MODEL = `${URL_PREFIX}-phi-4-mini/${VERSION_TAG}/original/phi-4-mini_bf16.pte`; +const PHI_4_MINI_4B_QUANTIZED_MODEL = `${URL_PREFIX}-phi-4-mini/${VERSION_TAG}/quantized/phi-4-mini_8da4w.pte`; +const PHI_4_MINI_TOKENIZER = `${URL_PREFIX}-phi-4-mini/${VERSION_TAG}/tokenizer.json`; +const PHI_4_MINI_TOKENIZER_CONFIG = `${URL_PREFIX}-phi-4-mini/${VERSION_TAG}/tokenizer_config.json`; + +export const PHI_4_MINI_4B = { + modelSource: PHI_4_MINI_4B_MODEL, + tokenizerSource: PHI_4_MINI_TOKENIZER, + tokenizerConfigSource: PHI_4_MINI_TOKENIZER_CONFIG, +}; + +export const PHI_4_MINI_4B_QUANTIZED = { + modelSource: PHI_4_MINI_4B_QUANTIZED_MODEL, + tokenizerSource: PHI_4_MINI_TOKENIZER, + tokenizerConfigSource: PHI_4_MINI_TOKENIZER_CONFIG, +}; // Classification -export const EFFICIENTNET_V2_S = +const EFFICIENTNET_V2_S_MODEL = Platform.OS === `ios` - ? `${BASE_URL_PREFIX}-efficientnet-v2-s/${VERSION_TAG}/coreml/efficientnet_v2_s_coreml_all.pte` - : `${BASE_URL_PREFIX}-efficientnet-v2-s/${VERSION_TAG}/xnnpack/efficientnet_v2_s_xnnpack.pte`; + ? `${URL_PREFIX}-efficientnet-v2-s/${VERSION_TAG}/coreml/efficientnet_v2_s_coreml_all.pte` + : `${URL_PREFIX}-efficientnet-v2-s/${VERSION_TAG}/xnnpack/efficientnet_v2_s_xnnpack.pte`; + +export const EFFICIENTNET_V2_S = { + modelSource: EFFICIENTNET_V2_S_MODEL, +}; // Object detection -export const SSDLITE_320_MOBILENET_V3_LARGE = `${BASE_URL_PREFIX}-ssdlite320-mobilenet-v3-large/${VERSION_TAG}/ssdlite320-mobilenetv3-large.pte`; +const SSDLITE_320_MOBILENET_V3_LARGE_MODEL = `${URL_PREFIX}-ssdlite320-mobilenet-v3-large/${VERSION_TAG}/ssdlite320-mobilenetv3-large.pte`; + +export const SSDLITE_320_MOBILENET_V3_LARGE = { + modelSource: SSDLITE_320_MOBILENET_V3_LARGE_MODEL, +}; // Style transfer -export const STYLE_TRANSFER_CANDY = +const STYLE_TRANSFER_CANDY_MODEL = Platform.OS === `ios` - ? `${BASE_URL_PREFIX}-style-transfer-candy/${VERSION_TAG}/coreml/style_transfer_candy_coreml.pte` - : `${BASE_URL_PREFIX}-style-transfer-candy/${VERSION_TAG}/xnnpack/style_transfer_candy_xnnpack.pte`; -export const STYLE_TRANSFER_MOSAIC = + ? `${URL_PREFIX}-style-transfer-candy/${VERSION_TAG}/coreml/style_transfer_candy_coreml.pte` + : `${URL_PREFIX}-style-transfer-candy/${VERSION_TAG}/xnnpack/style_transfer_candy_xnnpack.pte`; +const STYLE_TRANSFER_MOSAIC_MODEL = Platform.OS === `ios` - ? `${BASE_URL_PREFIX}-style-transfer-mosaic/${VERSION_TAG}/coreml/style_transfer_mosaic_coreml.pte` - : `${BASE_URL_PREFIX}-style-transfer-mosaic/${VERSION_TAG}/xnnpack/style_transfer_mosaic_xnnpack.pte`; -export const STYLE_TRANSFER_RAIN_PRINCESS = + ? `${URL_PREFIX}-style-transfer-mosaic/${VERSION_TAG}/coreml/style_transfer_mosaic_coreml.pte` + : `${URL_PREFIX}-style-transfer-mosaic/${VERSION_TAG}/xnnpack/style_transfer_mosaic_xnnpack.pte`; +const STYLE_TRANSFER_RAIN_PRINCESS_MODEL = Platform.OS === `ios` - ? `${BASE_URL_PREFIX}-style-transfer-rain-princess/${VERSION_TAG}/coreml/style_transfer_rain_princess_coreml.pte` - : `${BASE_URL_PREFIX}-style-transfer-rain-princess/${VERSION_TAG}/xnnpack/style_transfer_rain_princess_xnnpack.pte`; -export const STYLE_TRANSFER_UDNIE = + ? `${URL_PREFIX}-style-transfer-rain-princess/${VERSION_TAG}/coreml/style_transfer_rain_princess_coreml.pte` + : `${URL_PREFIX}-style-transfer-rain-princess/${VERSION_TAG}/xnnpack/style_transfer_rain_princess_xnnpack.pte`; +const STYLE_TRANSFER_UDNIE_MODEL = Platform.OS === `ios` - ? `${BASE_URL_PREFIX}-style-transfer-udnie/${VERSION_TAG}/coreml/style_transfer_udnie_coreml.pte` - : `${BASE_URL_PREFIX}-style-transfer-udnie/${VERSION_TAG}/xnnpack/style_transfer_udnie_xnnpack.pte`; + ? `${URL_PREFIX}-style-transfer-udnie/${VERSION_TAG}/coreml/style_transfer_udnie_coreml.pte` + : `${URL_PREFIX}-style-transfer-udnie/${VERSION_TAG}/xnnpack/style_transfer_udnie_xnnpack.pte`; + +export const STYLE_TRANSFER_CANDY = { + modelSource: STYLE_TRANSFER_CANDY_MODEL, +}; + +export const STYLE_TRANSFER_MOSAIC = { + modelSource: STYLE_TRANSFER_MOSAIC_MODEL, +}; + +export const STYLE_TRANSFER_RAIN_PRINCESS = { + modelSource: STYLE_TRANSFER_RAIN_PRINCESS_MODEL, +}; + +export const STYLE_TRANSFER_UDNIE = { + modelSource: STYLE_TRANSFER_UDNIE_MODEL, +}; // S2T -export const MOONSHINE_TINY_DECODER = `${BASE_URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_decoder.pte`; -export const MOONSHINE_TINY_ENCODER = `${BASE_URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_encoder.pte`; -export const MOONSHINE_TOKENIZER = `${BASE_URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/moonshine_tiny_tokenizer.json`; -export const WHISPER_TOKENIZER = `${BASE_URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/whisper_tokenizer.json`; -export const WHISPER_TINY_DECODER = `${BASE_URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_decoder.pte`; -export const WHISPER_TINY_ENCODER = `${BASE_URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_encoder.pte`; -export const WHISPER_TINY_MULTILINGUAL_ENCODER = `${BASE_URL_PREFIX}-whisper-tiny/${VERSION_TAG}/xnnpack/xnnpack_whisper_encoder.pte`; -export const WHISPER_TINY_MULTILINGUAL_DECODER = `${BASE_URL_PREFIX}-whisper-tiny/${VERSION_TAG}/xnnpack/xnnpack_whisper_decoder.pte`; -export const WHISPER_TINY_MULTILINGUAL_TOKENIZER = `${BASE_URL_PREFIX}-whisper-tiny/${VERSION_TAG}/tokenizer.json`; - -// OCR -export const DETECTOR_CRAFT_1280 = `${BASE_URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_1280.pte`; -export const DETECTOR_CRAFT_800 = `${BASE_URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_800.pte`; -export const DETECTOR_CRAFT_320 = `${BASE_URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_320.pte`; +const MOONSHINE_TINY_DECODER_MODEL = `${URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_decoder.pte`; +const MOONSHINE_TINY_ENCODER_MODEL = `${URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_encoder.pte`; +const MOONSHINE_TOKENIZER = `${URL_PREFIX}-moonshine-tiny/${VERSION_TAG}/moonshine_tiny_tokenizer.json`; +const WHISPER_TOKENIZER = `${URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/whisper_tokenizer.json`; +const WHISPER_TINY_DECODER_MODEL = `${URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_decoder.pte`; +const WHISPER_TINY_ENCODER_MODEL = `${URL_PREFIX}-whisper-tiny.en/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_encoder.pte`; +const WHISPER_TINY_MULTILINGUAL_ENCODER_MODEL = `${URL_PREFIX}-whisper-tiny/${VERSION_TAG}/xnnpack/xnnpack_whisper_encoder.pte`; +const WHISPER_TINY_MULTILINGUAL_DECODER_MODEL = `${URL_PREFIX}-whisper-tiny/${VERSION_TAG}/xnnpack/xnnpack_whisper_decoder.pte`; +const WHISPER_TINY_MULTILINGUAL_TOKENIZER = `${URL_PREFIX}-whisper-tiny/${VERSION_TAG}/tokenizer.json`; + +export const MOONSHINE_TINY = { + modelName: AvailableModels.MOONSHINE, + decoderSource: MOONSHINE_TINY_DECODER_MODEL, + encoderSource: MOONSHINE_TINY_ENCODER_MODEL, + tokenizerSource: MOONSHINE_TOKENIZER, +}; + +export const WHISPER_TINY = { + modelName: AvailableModels.WHISPER, + decoderSource: WHISPER_TINY_DECODER_MODEL, + encoderSource: WHISPER_TINY_ENCODER_MODEL, + tokenizerSource: WHISPER_TOKENIZER, +}; + +export const WHISPER_TINY_MULTILINGUAL = { + modelName: AvailableModels.WHISPER_MULTILINGUAL, + decoderSource: WHISPER_TINY_MULTILINGUAL_DECODER_MODEL, + encoderSource: WHISPER_TINY_MULTILINGUAL_ENCODER_MODEL, + tokenizerSource: WHISPER_TINY_MULTILINGUAL_TOKENIZER, +}; // Image segmentation -export const DEEPLAB_V3_RESNET50 = `${BASE_URL_PREFIX}-deeplab-v3/${VERSION_TAG}/xnnpack/deeplabV3_xnnpack_fp32.pte`; +const DEEPLAB_V3_RESNET50_MODEL = `${URL_PREFIX}-deeplab-v3/${VERSION_TAG}/xnnpack/deeplabV3_xnnpack_fp32.pte`; + +export const DEEPLAB_V3_RESNET50 = { + modelSource: DEEPLAB_V3_RESNET50_MODEL, +}; // Image Embeddings -export const CLIP_VIT_BASE_PATCH32_IMAGE_MODEL = - 'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32/resolve/v0.5.0/clip-vit-base-patch32-vision_xnnpack.pte'; +const CLIP_VIT_BASE_PATCH32_IMAGE_MODEL = `${URL_PREFIX}-clip-vit-base-patch32/${NEXT_VERSION_TAG}/clip-vit-base-patch32-vision_xnnpack.pte`; + +export const CLIP_VIT_BASE_PATCH32_IMAGE = { + modelSource: CLIP_VIT_BASE_PATCH32_IMAGE_MODEL, +}; // Text Embeddings -export const ALL_MINILM_L6_V2 = `${BASE_URL_PREFIX}-all-MiniLM-L6-v2/${VERSION_TAG}/all-MiniLM-L6-v2_xnnpack.pte`; -export const ALL_MINILM_L6_V2_TOKENIZER = `${BASE_URL_PREFIX}-all-MiniLM-L6-v2/${VERSION_TAG}/tokenizer.json`; +const ALL_MINILM_L6_V2_MODEL = `${URL_PREFIX}-all-MiniLM-L6-v2/${NEXT_VERSION_TAG}/all-MiniLM-L6-v2_xnnpack.pte`; +const ALL_MINILM_L6_V2_TOKENIZER = `${URL_PREFIX}-all-MiniLM-L6-v2/${NEXT_VERSION_TAG}/tokenizer.json`; +const ALL_MPNET_BASE_V2_MODEL = `${URL_PREFIX}-all-mpnet-base-v2/${NEXT_VERSION_TAG}/all-mpnet-base-v2_xnnpack.pte`; +const ALL_MPNET_BASE_V2_TOKENIZER = `${URL_PREFIX}-all-mpnet-base-v2/${NEXT_VERSION_TAG}/tokenizer.json`; +const MULTI_QA_MINILM_L6_COS_V1_MODEL = `${URL_PREFIX}-multi-qa-MiniLM-L6-cos-v1/${NEXT_VERSION_TAG}/multi-qa-MiniLM-L6-cos-v1_xnnpack.pte`; +const MULTI_QA_MINILM_L6_COS_V1_TOKENIZER = `${URL_PREFIX}-multi-qa-MiniLM-L6-cos-v1/${NEXT_VERSION_TAG}/tokenizer.json`; +const MULTI_QA_MPNET_BASE_DOT_V1_MODEL = `${URL_PREFIX}-multi-qa-mpnet-base-dot-v1/${NEXT_VERSION_TAG}/multi-qa-mpnet-base-dot-v1_xnnpack.pte`; +const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER = `${URL_PREFIX}-multi-qa-mpnet-base-dot-v1/${NEXT_VERSION_TAG}/tokenizer.json`; +const CLIP_VIT_BASE_PATCH32_TEXT_MODEL = `${URL_PREFIX}-clip-vit-base-patch32/${NEXT_VERSION_TAG}/clip-vit-base-patch32-text_xnnpack.pte`; +const CLIP_VIT_BASE_PATCH32_TEXT_TOKENIZER = `${URL_PREFIX}-clip-vit-base-patch32/${NEXT_VERSION_TAG}/tokenizer.json`; -export const ALL_MPNET_BASE_V2 = `${BASE_URL_PREFIX}-all-mpnet-base-v2/${VERSION_TAG}/all-mpnet-base-v2_xnnpack.pte`; -export const ALL_MPNET_BASE_V2_TOKENIZER = `${BASE_URL_PREFIX}-all-mpnet-base-v2/${VERSION_TAG}/tokenizer.json`; +export const ALL_MINILM_L6_V2 = { + modelSource: ALL_MINILM_L6_V2_MODEL, + tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER, +}; -export const MULTI_QA_MINILM_L6_COS_V1 = `${BASE_URL_PREFIX}-multi-qa-MiniLM-L6-cos-v1/${VERSION_TAG}/multi-qa-MiniLM-L6-cos-v1_xnnpack.pte`; -export const MULTI_QA_MINILM_L6_COS_V1_TOKENIZER = `${BASE_URL_PREFIX}-multi-qa-MiniLM-L6-cos-v1/${VERSION_TAG}/tokenizer.json`; +export const ALL_MPNET_BASE_V2 = { + modelSource: ALL_MPNET_BASE_V2_MODEL, + tokenizerSource: ALL_MPNET_BASE_V2_TOKENIZER, +}; -export const MULTI_QA_MPNET_BASE_DOT_V1 = `${BASE_URL_PREFIX}-multi-qa-mpnet-base-dot-v1/${VERSION_TAG}/multi-qa-mpnet-base-dot-v1_xnnpack.pte`; -export const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER = `${BASE_URL_PREFIX}-multi-qa-mpnet-base-dot-v1/${VERSION_TAG}/tokenizer.json`; +export const MULTI_QA_MINILM_L6_COS_V1 = { + modelSource: MULTI_QA_MINILM_L6_COS_V1_MODEL, + tokenizerSource: MULTI_QA_MINILM_L6_COS_V1_TOKENIZER, +}; -export const CLIP_VIT_BASE_PATCH32_TEXT_MODEL = - 'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32/resolve/v0.5.0/clip-vit-base-patch32-text_xnnpack.pte'; -export const CLIP_VIT_BASE_PATCH32_TEXT_TOKENIZER = - 'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32/resolve/v0.5.0/tokenizer.json'; +export const MULTI_QA_MPNET_BASE_DOT_V1 = { + modelSource: MULTI_QA_MPNET_BASE_DOT_V1_MODEL, + tokenizerSource: MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER, +}; export const CLIP_VIT_BASE_PATCH32_TEXT = { modelSource: CLIP_VIT_BASE_PATCH32_TEXT_MODEL, tokenizerSource: CLIP_VIT_BASE_PATCH32_TEXT_TOKENIZER, }; - -export const CLIP_VIT_BASE_PATCH32_IMAGE = { - modelSource: CLIP_VIT_BASE_PATCH32_IMAGE_MODEL, -}; - - -// Backward compatibility -export const LLAMA3_2_3B_URL = LLAMA3_2_3B; -export const LLAMA3_2_3B_QLORA_URL = LLAMA3_2_3B_QLORA; -export const LLAMA3_2_3B_SPINQUANT_URL = LLAMA3_2_3B_SPINQUANT; -export const LLAMA3_2_1B_URL = LLAMA3_2_1B; -export const LLAMA3_2_1B_QLORA_URL = LLAMA3_2_1B_QLORA; -export const LLAMA3_2_1B_SPINQUANT_URL = LLAMA3_2_1B_SPINQUANT; -export const LLAMA3_2_1B_TOKENIZER = LLAMA3_2_TOKENIZER; -export const LLAMA3_2_3B_TOKENIZER = LLAMA3_2_TOKENIZER; diff --git a/packages/react-native-executorch/src/constants/ocr/models.ts b/packages/react-native-executorch/src/constants/ocr/models.ts index 4f198612c0..a5bef1b181 100644 --- a/packages/react-native-executorch/src/constants/ocr/models.ts +++ b/packages/react-native-executorch/src/constants/ocr/models.ts @@ -1,453 +1,885 @@ -const createHFRecognizerDownloadUrl = (alphabet: string, size: number) => { - return `https://huggingface.co/software-mansion/react-native-executorch-recognizer-crnn.en/resolve/v0.4.0/xnnpack/${alphabet}/xnnpack_crnn_${alphabet}_${size}.pte`; -}; -export const RECOGNIZER_ENGLISH_CRNN_512 = createHFRecognizerDownloadUrl( +import { symbols } from './symbols'; + +const URL_PREFIX = + 'https://huggingface.co/software-mansion/react-native-executorch'; +const VERSION_TAG = 'resolve/v0.4.0'; + +const DETECTOR_CRAFT_1280_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_1280.pte`; +const DETECTOR_CRAFT_800_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_800.pte`; +const DETECTOR_CRAFT_320_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_320.pte`; + +const createHFRecognizerDownloadUrl = (alphabet: string, size: number) => + `${URL_PREFIX}-recognizer-crnn.en/${VERSION_TAG}/xnnpack/${alphabet}/xnnpack_crnn_${alphabet}_${size}.pte`; + +const RECOGNIZER_ENGLISH_CRNN_512 = createHFRecognizerDownloadUrl( 'english', 512 ); -export const RECOGNIZER_ENGLISH_CRNN_256 = createHFRecognizerDownloadUrl( +const RECOGNIZER_ENGLISH_CRNN_256 = createHFRecognizerDownloadUrl( 'english', 256 ); -export const RECOGNIZER_ENGLISH_CRNN_128 = createHFRecognizerDownloadUrl( +const RECOGNIZER_ENGLISH_CRNN_128 = createHFRecognizerDownloadUrl( 'english', 128 ); -export const RECOGNIZER_ENGLISH_CRNN_64 = createHFRecognizerDownloadUrl( - 'english', - 64 -); +const RECOGNIZER_ENGLISH_CRNN_64 = createHFRecognizerDownloadUrl('english', 64); -export const RECOGNIZER_LATIN_CRNN_512 = createHFRecognizerDownloadUrl( - 'latin', - 512 -); -export const RECOGNIZER_LATIN_CRNN_256 = createHFRecognizerDownloadUrl( - 'latin', - 256 -); -export const RECOGNIZER_LATIN_CRNN_128 = createHFRecognizerDownloadUrl( - 'latin', - 128 -); -export const RECOGNIZER_LATIN_CRNN_64 = createHFRecognizerDownloadUrl( - 'latin', - 64 -); +const RECOGNIZER_LATIN_CRNN_512 = createHFRecognizerDownloadUrl('latin', 512); +const RECOGNIZER_LATIN_CRNN_256 = createHFRecognizerDownloadUrl('latin', 256); +const RECOGNIZER_LATIN_CRNN_128 = createHFRecognizerDownloadUrl('latin', 128); +const RECOGNIZER_LATIN_CRNN_64 = createHFRecognizerDownloadUrl('latin', 64); -export const RECOGNIZER_JAPANESE_CRNN_512 = createHFRecognizerDownloadUrl( +const RECOGNIZER_JAPANESE_CRNN_512 = createHFRecognizerDownloadUrl( 'japanese', 512 ); -export const RECOGNIZER_JAPANESE_CRNN_256 = createHFRecognizerDownloadUrl( +const RECOGNIZER_JAPANESE_CRNN_256 = createHFRecognizerDownloadUrl( 'japanese', 256 ); -export const RECOGNIZER_JAPANESE_CRNN_128 = createHFRecognizerDownloadUrl( +const RECOGNIZER_JAPANESE_CRNN_128 = createHFRecognizerDownloadUrl( 'japanese', 128 ); -export const RECOGNIZER_JAPANESE_CRNN_64 = createHFRecognizerDownloadUrl( +const RECOGNIZER_JAPANESE_CRNN_64 = createHFRecognizerDownloadUrl( 'japanese', 64 ); -export const RECOGNIZER_KANNADA_CRNN_512 = createHFRecognizerDownloadUrl( +const RECOGNIZER_KANNADA_CRNN_512 = createHFRecognizerDownloadUrl( 'kannada', 512 ); -export const RECOGNIZER_KANNADA_CRNN_256 = createHFRecognizerDownloadUrl( +const RECOGNIZER_KANNADA_CRNN_256 = createHFRecognizerDownloadUrl( 'kannada', 256 ); -export const RECOGNIZER_KANNADA_CRNN_128 = createHFRecognizerDownloadUrl( +const RECOGNIZER_KANNADA_CRNN_128 = createHFRecognizerDownloadUrl( 'kannada', 128 ); -export const RECOGNIZER_KANNADA_CRNN_64 = createHFRecognizerDownloadUrl( - 'kannada', - 64 -); +const RECOGNIZER_KANNADA_CRNN_64 = createHFRecognizerDownloadUrl('kannada', 64); + +const RECOGNIZER_KOREAN_CRNN_512 = createHFRecognizerDownloadUrl('korean', 512); +const RECOGNIZER_KOREAN_CRNN_256 = createHFRecognizerDownloadUrl('korean', 256); +const RECOGNIZER_KOREAN_CRNN_128 = createHFRecognizerDownloadUrl('korean', 128); +const RECOGNIZER_KOREAN_CRNN_64 = createHFRecognizerDownloadUrl('korean', 64); + +const RECOGNIZER_TELUGU_CRNN_512 = createHFRecognizerDownloadUrl('telugu', 512); +const RECOGNIZER_TELUGU_CRNN_256 = createHFRecognizerDownloadUrl('telugu', 256); +const RECOGNIZER_TELUGU_CRNN_128 = createHFRecognizerDownloadUrl('telugu', 128); +const RECOGNIZER_TELUGU_CRNN_64 = createHFRecognizerDownloadUrl('telugu', 64); + +const RECOGNIZER_ZH_SIM_CRNN_512 = createHFRecognizerDownloadUrl('zh-sim', 512); +const RECOGNIZER_ZH_SIM_CRNN_256 = createHFRecognizerDownloadUrl('zh-sim', 256); +const RECOGNIZER_ZH_SIM_CRNN_128 = createHFRecognizerDownloadUrl('zh-sim', 128); +const RECOGNIZER_ZH_SIM_CRNN_64 = createHFRecognizerDownloadUrl('zh-sim', 64); -export const RECOGNIZER_KOREAN_CRNN_512 = createHFRecognizerDownloadUrl( - 'korean', +const RECOGNIZER_CYRILLIC_CRNN_512 = createHFRecognizerDownloadUrl( + 'cyrillic', 512 ); -export const RECOGNIZER_KOREAN_CRNN_256 = createHFRecognizerDownloadUrl( - 'korean', +const RECOGNIZER_CYRILLIC_CRNN_256 = createHFRecognizerDownloadUrl( + 'cyrillic', 256 ); -export const RECOGNIZER_KOREAN_CRNN_128 = createHFRecognizerDownloadUrl( - 'korean', +const RECOGNIZER_CYRILLIC_CRNN_128 = createHFRecognizerDownloadUrl( + 'cyrillic', 128 ); -export const RECOGNIZER_KOREAN_CRNN_64 = createHFRecognizerDownloadUrl( - 'korean', +const RECOGNIZER_CYRILLIC_CRNN_64 = createHFRecognizerDownloadUrl( + 'cyrillic', 64 ); -export const RECOGNIZER_TELUGU_CRNN_512 = createHFRecognizerDownloadUrl( - 'telugu', - 512 +const createOCRObject = ( + recognizerLarge: string, + recognizerMedium: string, + recognizerSmall: string, + language: keyof typeof symbols +) => { + return { + detectorSource: DETECTOR_CRAFT_800_MODEL, + recognizerSources: { + recognizerLarge, + recognizerMedium, + recognizerSmall, + }, + language, + }; +}; + +const createVerticalOCRObject = ( + recognizerLarge: string, + recognizerSmall: string, + language: keyof typeof symbols +) => { + return { + detectorSources: { + detectorLarge: DETECTOR_CRAFT_1280_MODEL, + detectorNarrow: DETECTOR_CRAFT_320_MODEL, + }, + recognizerSources: { + recognizerLarge, + recognizerSmall, + }, + language, + }; +}; + +export const OCR_ABAZA = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'abq' ); -export const RECOGNIZER_TELUGU_CRNN_256 = createHFRecognizerDownloadUrl( - 'telugu', - 256 +export const VERTICAL_OCR_ABAZA = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'abq' ); -export const RECOGNIZER_TELUGU_CRNN_128 = createHFRecognizerDownloadUrl( - 'telugu', - 128 + +export const OCR_ADYGHE = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'ady' ); -export const RECOGNIZER_TELUGU_CRNN_64 = createHFRecognizerDownloadUrl( - 'telugu', - 64 +export const VERTICAL_OCR_ADYGHE = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'ady' ); -export const RECOGNIZER_ZH_SIM_CRNN_512 = createHFRecognizerDownloadUrl( - 'zh-sim', - 512 +export const OCR_AFRIKAANS = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'af' ); -export const RECOGNIZER_ZH_SIM_CRNN_256 = createHFRecognizerDownloadUrl( - 'zh-sim', - 256 +export const VERTICAL_OCR_AFRIKAANS = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'af' ); -export const RECOGNIZER_ZH_SIM_CRNN_128 = createHFRecognizerDownloadUrl( - 'zh-sim', - 128 + +export const OCR_AVAR = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'ava' ); -export const RECOGNIZER_ZH_SIM_CRNN_64 = createHFRecognizerDownloadUrl( - 'zh-sim', - 64 +export const VERTICAL_OCR_AVAR = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'ava' ); -export const RECOGNIZER_CYRILLIC_CRNN_512 = createHFRecognizerDownloadUrl( - 'cyrillic', - 512 +export const OCR_AZERBAIJANI = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'az' ); -export const RECOGNIZER_CYRILLIC_CRNN_256 = createHFRecognizerDownloadUrl( - 'cyrillic', - 256 +export const VERTICAL_OCR_AZERBAIJANI = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'az' ); -export const RECOGNIZER_CYRILLIC_CRNN_128 = createHFRecognizerDownloadUrl( - 'cyrillic', - 128 + +export const OCR_BELARUSIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'be' ); -export const RECOGNIZER_CYRILLIC_CRNN_64 = createHFRecognizerDownloadUrl( - 'cyrillic', - 64 +export const VERTICAL_OCR_BELARUSIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'be' +); + +export const OCR_BULGARIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'bg' +); +export const VERTICAL_OCR_BULGARIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'bg' +); + +export const OCR_BOSNIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'bs' +); +export const VERTICAL_OCR_BOSNIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'bs' +); + +export const OCR_SIMPLIFIED_CHINESE = createOCRObject( + RECOGNIZER_ZH_SIM_CRNN_512, + RECOGNIZER_ZH_SIM_CRNN_256, + RECOGNIZER_ZH_SIM_CRNN_128, + 'chSim' +); +export const VERTICAL_OCR_SIMPLIFIED_CHINESE = createVerticalOCRObject( + RECOGNIZER_ZH_SIM_CRNN_512, + RECOGNIZER_ZH_SIM_CRNN_64, + 'chSim' +); + +export const OCR_CHECHEN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'che' +); +export const VERTICAL_OCR_CHECHEN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'che' +); + +export const OCR_CZECH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'cs' +); +export const VERTICAL_OCR_CZECH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'cs' +); + +export const OCR_WELSH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'cy' +); +export const VERTICAL_OCR_WELSH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'cy' +); + +export const OCR_DANISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'da' +); +export const VERTICAL_OCR_DANISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'da' +); + +export const OCR_DARGWA = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'dar' +); +export const VERTICAL_OCR_DARGWA = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'dar' +); + +export const OCR_GERMAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'de' +); +export const VERTICAL_OCR_GERMAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'de' +); + +export const OCR_ENGLISH = createOCRObject( + RECOGNIZER_ENGLISH_CRNN_512, + RECOGNIZER_ENGLISH_CRNN_256, + RECOGNIZER_ENGLISH_CRNN_128, + 'en' +); +export const VERTICAL_OCR_ENGLISH = createVerticalOCRObject( + RECOGNIZER_ENGLISH_CRNN_512, + RECOGNIZER_ENGLISH_CRNN_64, + 'en' +); + +export const OCR_SPANISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'es' +); +export const VERTICAL_OCR_SPANISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'es' +); + +export const OCR_ESTONIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'et' +); +export const VERTICAL_OCR_ESTONIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'et' +); + +export const OCR_FRENCH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'fr' +); +export const VERTICAL_OCR_FRENCH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'fr' +); + +export const OCR_IRISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'ga' +); +export const VERTICAL_OCR_IRISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'ga' +); + +export const OCR_CROATIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'hr' +); +export const VERTICAL_OCR_CROATIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'hr' +); + +export const OCR_HUNGARIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'hu' +); +export const VERTICAL_OCR_HUNGARIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'hu' +); + +export const OCR_INDONESIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'id' +); +export const VERTICAL_OCR_INDONESIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'id' +); + +export const OCR_INGUSH = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'inh' +); +export const VERTICAL_OCR_INGUSH = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'inh' +); + +export const OCR_ICELANDIC = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'ic' +); +export const VERTICAL_OCR_ICELANDIC = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'ic' +); + +export const OCR_ITALIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'it' +); +export const VERTICAL_OCR_ITALIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'it' +); + +export const OCR_JAPANESE = createOCRObject( + RECOGNIZER_JAPANESE_CRNN_512, + RECOGNIZER_JAPANESE_CRNN_256, + RECOGNIZER_JAPANESE_CRNN_128, + 'ja' +); +export const VERTICAL_OCR_JAPANESE = createVerticalOCRObject( + RECOGNIZER_JAPANESE_CRNN_512, + RECOGNIZER_JAPANESE_CRNN_64, + 'ja' +); + +export const OCR_KARBADIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'kbd' +); +export const VERTICAL_OCR_KARBADIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'kbd' +); + +export const OCR_KANNADA = createOCRObject( + RECOGNIZER_KANNADA_CRNN_512, + RECOGNIZER_KANNADA_CRNN_256, + RECOGNIZER_KANNADA_CRNN_128, + 'kn' +); +export const VERTICAL_OCR_KANNADA = createVerticalOCRObject( + RECOGNIZER_KANNADA_CRNN_512, + RECOGNIZER_KANNADA_CRNN_64, + 'kn' +); + +export const OCR_KOREAN = createOCRObject( + RECOGNIZER_KOREAN_CRNN_512, + RECOGNIZER_KOREAN_CRNN_256, + RECOGNIZER_KOREAN_CRNN_128, + 'ko' +); +export const VERTICAL_OCR_KOREAN = createVerticalOCRObject( + RECOGNIZER_KOREAN_CRNN_512, + RECOGNIZER_KOREAN_CRNN_64, + 'ko' +); + +export const OCR_KURDISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'ku' +); +export const VERTICAL_OCR_KURDISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'ku' +); + +export const OCR_LATIN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'la' +); +export const VERTICAL_OCR_LATIN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'la' +); + +export const OCR_LAK = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'lbe' +); +export const VERTICAL_OCR_LAK = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'lbe' +); + +export const OCR_LEZGHIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'lez' +); +export const VERTICAL_OCR_LEZGHIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'lez' +); + +export const OCR_LITHUANIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'lt' +); +export const VERTICAL_OCR_LITHUANIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'lt' +); + +export const OCR_LATVIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'lv' +); +export const VERTICAL_OCR_LATVIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'lv' +); + +export const OCR_MAORI = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'mi' +); +export const VERTICAL_OCR_MAORI = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'mi' +); + +export const OCR_MONGOLIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'mn' +); +export const VERTICAL_OCR_MONGOLIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'mn' +); + +export const OCR_MALAY = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'ms' +); +export const VERTICAL_OCR_MALAY = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'ms' +); + +export const OCR_MALTESE = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'mt' +); +export const VERTICAL_OCR_MALTESE = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'mt' +); + +export const OCR_DUTCH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'nl' +); +export const VERTICAL_OCR_DUTCH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'nl' +); + +export const OCR_NORWEGIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'no' +); +export const VERTICAL_OCR_NORWEGIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'no' +); + +export const OCR_OCCITAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'oc' +); +export const VERTICAL_OCR_OCCITAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'oc' +); + +export const OCR_PALI = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'pi' +); +export const VERTICAL_OCR_PALI = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'pi' +); + +export const OCR_POLISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'pl' +); +export const VERTICAL_OCR_POLISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'pl' ); -export const RECOGNIZER_ABQ_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_ABQ_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_ABQ_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_ABQ_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_ADY_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_ADY_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_ADY_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_ADY_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_AF_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_AF_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_AF_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_AF_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_AVA_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_AVA_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_AVA_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_AVA_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_AZ_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_AZ_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_AZ_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_AZ_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_BE_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_BE_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_BE_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_BE_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_BG_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_BG_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_BG_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_BG_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_BS_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_BS_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_BS_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_BS_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_CH_SIM_CRNN_512 = RECOGNIZER_ZH_SIM_CRNN_512; -export const RECOGNIZER_CH_SIM_CRNN_256 = RECOGNIZER_ZH_SIM_CRNN_256; -export const RECOGNIZER_CH_SIM_CRNN_128 = RECOGNIZER_ZH_SIM_CRNN_128; -export const RECOGNIZER_CH_SIM_CRNN_64 = RECOGNIZER_ZH_SIM_CRNN_64; - -export const RECOGNIZER_CHE_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_CHE_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_CHE_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_CHE_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_CS_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_CS_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_CS_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_CS_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_CY_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_CY_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_CY_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_CY_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_DA_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_DA_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_DA_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_DA_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_DAR_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_DAR_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_DAR_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_DAR_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_DE_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_DE_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_DE_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_DE_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_EN_CRNN_512 = RECOGNIZER_ENGLISH_CRNN_512; -export const RECOGNIZER_EN_CRNN_256 = RECOGNIZER_ENGLISH_CRNN_256; -export const RECOGNIZER_EN_CRNN_128 = RECOGNIZER_ENGLISH_CRNN_128; -export const RECOGNIZER_EN_CRNN_64 = RECOGNIZER_ENGLISH_CRNN_64; - -export const RECOGNIZER_ES_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_ES_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_ES_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_ES_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_ET_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_ET_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_ET_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_ET_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_FR_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_FR_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_FR_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_FR_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_GA_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_GA_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_GA_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_GA_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_HR_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_HR_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_HR_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_HR_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_HU_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_HU_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_HU_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_HU_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_ID_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_ID_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_ID_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_ID_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_INH_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_INH_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_INH_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_INH_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_IC_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_IC_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_IC_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_IC_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_IT_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_IT_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_IT_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_IT_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_JA_CRNN_512 = RECOGNIZER_JAPANESE_CRNN_512; -export const RECOGNIZER_JA_CRNN_256 = RECOGNIZER_JAPANESE_CRNN_256; -export const RECOGNIZER_JA_CRNN_128 = RECOGNIZER_JAPANESE_CRNN_128; -export const RECOGNIZER_JA_CRNN_64 = RECOGNIZER_JAPANESE_CRNN_64; - -export const RECOGNIZER_KBD_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_KBD_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_KBD_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_KBD_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_KN_CRNN_512 = RECOGNIZER_KANNADA_CRNN_512; -export const RECOGNIZER_KN_CRNN_256 = RECOGNIZER_KANNADA_CRNN_256; -export const RECOGNIZER_KN_CRNN_128 = RECOGNIZER_KANNADA_CRNN_128; -export const RECOGNIZER_KN_CRNN_64 = RECOGNIZER_KANNADA_CRNN_64; - -export const RECOGNIZER_KO_CRNN_512 = RECOGNIZER_KOREAN_CRNN_512; -export const RECOGNIZER_KO_CRNN_256 = RECOGNIZER_KOREAN_CRNN_256; -export const RECOGNIZER_KO_CRNN_128 = RECOGNIZER_KOREAN_CRNN_128; -export const RECOGNIZER_KO_CRNN_64 = RECOGNIZER_KOREAN_CRNN_64; - -export const RECOGNIZER_KU_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_KU_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_KU_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_KU_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_LA_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_LA_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_LA_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_LA_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_LBE_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_LBE_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_LBE_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_LBE_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_LEZ_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_LEZ_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_LEZ_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_LEZ_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_LT_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_LT_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_LT_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_LT_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_LV_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_LV_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_LV_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_LV_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_MI_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_MI_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_MI_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_MI_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_MN_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_MN_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_MN_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_MN_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_MS_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_MS_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_MS_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_MS_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_MT_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_MT_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_MT_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_MT_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_NL_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_NL_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_NL_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_NL_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_NO_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_NO_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_NO_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_NO_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_OC_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_OC_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_OC_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_OC_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_PI_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_PI_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_PI_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_PI_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_PL_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_PL_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_PL_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_PL_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_PT_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_PT_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_PT_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_PT_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_RO_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_RO_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_RO_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_RO_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_RU_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_RU_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_RU_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_RU_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_RS_CYRILLIC_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_RS_CYRILLIC_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_RS_CYRILLIC_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_RS_CYRILLIC_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_RS_LATIN_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_RS_LATIN_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_RS_LATIN_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_RS_LATIN_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_SK_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_SK_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_SK_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_SK_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_SL_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_SL_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_SL_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_SL_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_SQ_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_SQ_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_SQ_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_SQ_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_SV_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_SV_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_SV_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_SV_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_SW_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_SW_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_SW_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_SW_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_TAB_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_TAB_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_TAB_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_TAB_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_TE_CRNN_512 = RECOGNIZER_TELUGU_CRNN_512; -export const RECOGNIZER_TE_CRNN_256 = RECOGNIZER_TELUGU_CRNN_256; -export const RECOGNIZER_TE_CRNN_128 = RECOGNIZER_TELUGU_CRNN_128; -export const RECOGNIZER_TE_CRNN_64 = RECOGNIZER_TELUGU_CRNN_64; - -export const RECOGNIZER_TJK_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_TJK_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_TJK_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_TJK_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_TL_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_TL_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_TL_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_TL_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_TR_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_TR_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_TR_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_TR_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_UK_CRNN_512 = RECOGNIZER_CYRILLIC_CRNN_512; -export const RECOGNIZER_UK_CRNN_256 = RECOGNIZER_CYRILLIC_CRNN_256; -export const RECOGNIZER_UK_CRNN_128 = RECOGNIZER_CYRILLIC_CRNN_128; -export const RECOGNIZER_UK_CRNN_64 = RECOGNIZER_CYRILLIC_CRNN_64; - -export const RECOGNIZER_UZ_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_UZ_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_UZ_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_UZ_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; - -export const RECOGNIZER_VI_CRNN_512 = RECOGNIZER_LATIN_CRNN_512; -export const RECOGNIZER_VI_CRNN_256 = RECOGNIZER_LATIN_CRNN_256; -export const RECOGNIZER_VI_CRNN_128 = RECOGNIZER_LATIN_CRNN_128; -export const RECOGNIZER_VI_CRNN_64 = RECOGNIZER_LATIN_CRNN_64; +export const OCR_PORTUGUESE = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'pt' +); +export const VERTICAL_OCR_PORTUGUESE = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'pt' +); + +export const OCR_ROMANIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'ro' +); +export const VERTICAL_OCR_ROMANIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'ro' +); + +export const OCR_RUSSIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'ru' +); +export const VERTICAL_OCR_RUSSIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'ru' +); + +export const OCR_SERBIAN_CYRILLIC = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'rsCyrillic' +); +export const VERTICAL_OCR_SERBIAN_CYRILLIC = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'rsCyrillic' +); + +export const OCR_SERBIAN_LATIN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'rsLatin' +); +export const VERTICAL_OCR_SERBIAN_LATIN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'rsLatin' +); + +export const OCR_SLOVAK = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'sk' +); +export const VERTICAL_OCR_SLOVAK = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'sk' +); + +export const OCR_SLOVENIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'sl' +); +export const VERTICAL_OCR_SLOVENIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'sl' +); + +export const OCR_ALBANIAN = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'sq' +); +export const VERTICAL_OCR_ALBANIAN = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'sq' +); + +export const OCR_SWEDISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'sv' +); +export const VERTICAL_OCR_SWEDISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'sv' +); + +export const OCR_SWAHILI = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'sw' +); +export const VERTICAL_OCR_SWAHILI = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'sw' +); + +export const OCR_TABASSARAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'tab' +); +export const VERTICAL_OCR_TABASSARAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'tab' +); + +export const OCR_TELUGU = createOCRObject( + RECOGNIZER_TELUGU_CRNN_512, + RECOGNIZER_TELUGU_CRNN_256, + RECOGNIZER_TELUGU_CRNN_128, + 'te' +); +export const VERTICAL_OCR_TELUGU = createVerticalOCRObject( + RECOGNIZER_TELUGU_CRNN_512, + RECOGNIZER_TELUGU_CRNN_64, + 'te' +); + +export const OCR_TAJIK = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'tjk' +); +export const VERTICAL_OCR_TAJIK = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'tjk' +); + +export const OCR_TAGALOG = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'tl' +); +export const VERTICAL_OCR_TAGALOG = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'tl' +); + +export const OCR_TURKISH = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'tr' +); +export const VERTICAL_OCR_TURKISH = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'tr' +); + +export const OCR_UKRAINIAN = createOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_256, + RECOGNIZER_CYRILLIC_CRNN_128, + 'uk' +); +export const VERTICAL_OCR_UKRAINIAN = createVerticalOCRObject( + RECOGNIZER_CYRILLIC_CRNN_512, + RECOGNIZER_CYRILLIC_CRNN_64, + 'uk' +); + +export const OCR_UZBEK = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'uz' +); +export const VERTICAL_OCR_UZBEK = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'uz' +); + +export const OCR_VIETNAMESE = createOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_256, + RECOGNIZER_LATIN_CRNN_128, + 'vi' +); +export const VERTICAL_OCR_VIETNAMESE = createVerticalOCRObject( + RECOGNIZER_LATIN_CRNN_512, + RECOGNIZER_LATIN_CRNN_64, + 'vi' +); diff --git a/packages/react-native-executorch/src/constants/ocr/symbols.ts b/packages/react-native-executorch/src/constants/ocr/symbols.ts index 0bbc92ac5c..99471c4709 100644 --- a/packages/react-native-executorch/src/constants/ocr/symbols.ts +++ b/packages/react-native-executorch/src/constants/ocr/symbols.ts @@ -19,130 +19,130 @@ export const alphabets = { }; export const symbols = { - //Abaza + // Abaza abq: alphabets.cyrillic, - //Adyghe + // Adyghe ady: alphabets.cyrillic, - //Africans + // Africans af: alphabets.latin, - //Avar + // Avar ava: alphabets.cyrillic, - //Azerbaijani + // Azerbaijani az: alphabets.latin, - //Belarusian + // Belarusian be: alphabets.cyrillic, - //Bulgarian + // Bulgarian bg: alphabets.cyrillic, - //Bosnian + // Bosnian bs: alphabets.latin, - //Simplified Chinese + // Simplified Chinese chSim: alphabets.zhSim, - //Chechen + // Chechen che: alphabets.cyrillic, - //Czech + // Czech cs: alphabets.latin, - //Welsh + // Welsh cy: alphabets.latin, - //Danish + // Danish da: alphabets.latin, - //Dargwa + // Dargwa dar: alphabets.cyrillic, - //German + // German de: alphabets.latin, - //English + // English en: alphabets.english, - //Spanish + // Spanish es: alphabets.latin, - //Estonian + // Estonian et: alphabets.latin, - //French + // French fr: alphabets.latin, - //Irish + // Irish ga: alphabets.latin, - //Croatian + // Croatian hr: alphabets.latin, - //Hungarian + // Hungarian hu: alphabets.latin, - //Indonesian + // Indonesian id: alphabets.latin, - //Ingush + // Ingush inh: alphabets.cyrillic, - //Icelandic + // Icelandic ic: alphabets.latin, - //Italian + // Italian it: alphabets.latin, - //Japanese + // Japanese ja: alphabets.japanese, - //Karbadian + // Karbadian kbd: alphabets.cyrillic, - //Kannada + // Kannada kn: alphabets.kannada, - //Korean + // Korean ko: alphabets.korean, - //Kurdish + // Kurdish ku: alphabets.latin, - //Latin + // Latin la: alphabets.latin, - //Lak + // Lak lbe: alphabets.cyrillic, - //Lezghian + // Lezghian lez: alphabets.cyrillic, - //Lithuanian + // Lithuanian lt: alphabets.latin, - //Latvian + // Latvian lv: alphabets.latin, - //Maori + // Maori mi: alphabets.latin, - //Mongolian + // Mongolian mn: alphabets.cyrillic, - //Malay + // Malay ms: alphabets.latin, - //Maltese + // Maltese mt: alphabets.latin, - //Dutch + // Dutch nl: alphabets.latin, - //Norwegian + // Norwegian no: alphabets.latin, - //Occitan + // Occitan oc: alphabets.latin, - //Pali + // Pali pi: alphabets.latin, - //Polish + // Polish pl: alphabets.latin, - //Portuguese + // Portuguese pt: alphabets.latin, - //Romanian + // Romanian ro: alphabets.latin, - //Russian + // Russian ru: alphabets.cyrillic, - //Serbian (cyrillic) + // Serbian (cyrillic) rsCyrillic: alphabets.cyrillic, - //Serbian (latin) + // Serbian (latin) rsLatin: alphabets.latin, - //Slovak + // Slovak sk: alphabets.latin, - //Slovenian + // Slovenian sl: alphabets.latin, - //Albanian + // Albanian sq: alphabets.latin, - //Swedish + // Swedish sv: alphabets.latin, - //Swahili + // Swahili sw: alphabets.latin, - //Tabassaran + // Tabassaran tab: alphabets.cyrillic, - //Telugu + // Telugu te: alphabets.telugu, - //Tajik + // Tajik tjk: alphabets.cyrillic, - //Tagalog + // Tagalog tl: alphabets.latin, - //Turkish: + // Turkish tr: alphabets.latin, - //Ukrainian + // Ukrainian uk: alphabets.cyrillic, - //Uzbek + // Uzbek uz: alphabets.latin, - //Vietnamese + // Vietnamese vi: alphabets.latin, }; diff --git a/packages/react-native-executorch/src/constants/sttDefaults.ts b/packages/react-native-executorch/src/constants/sttDefaults.ts index 647c9a1468..655e37b6ff 100644 --- a/packages/react-native-executorch/src/constants/sttDefaults.ts +++ b/packages/react-native-executorch/src/constants/sttDefaults.ts @@ -1,13 +1,7 @@ import { - MOONSHINE_TINY_ENCODER, - MOONSHINE_TINY_DECODER, - MOONSHINE_TOKENIZER, - WHISPER_TINY_ENCODER, - WHISPER_TINY_DECODER, - WHISPER_TOKENIZER, - WHISPER_TINY_MULTILINGUAL_ENCODER, - WHISPER_TINY_MULTILINGUAL_DECODER, - WHISPER_TINY_MULTILINGUAL_TOKENIZER, + MOONSHINE_TINY, + WHISPER_TINY, + WHISPER_TINY_MULTILINGUAL, } from './modelUrls'; import { AvailableModels, ModelConfig } from '../types/stt'; @@ -17,11 +11,11 @@ export const HAMMING_DIST_THRESHOLD = 1; const whisperTinyModelConfig = { sources: { - encoder: WHISPER_TINY_ENCODER, - decoder: WHISPER_TINY_DECODER, + encoder: WHISPER_TINY.encoderSource, + decoder: WHISPER_TINY.decoderSource, }, tokenizer: { - source: WHISPER_TOKENIZER, + source: WHISPER_TINY.tokenizerSource, bos: 50257, // FIXME: this is a placeholder and needs to be changed eos: 50256, // FIXME: this is a placeholder and needs to be changed }, @@ -30,11 +24,11 @@ const whisperTinyModelConfig = { const moonshineTinyModelConfig = { sources: { - encoder: MOONSHINE_TINY_ENCODER, - decoder: MOONSHINE_TINY_DECODER, + encoder: MOONSHINE_TINY.encoderSource, + decoder: MOONSHINE_TINY.decoderSource, }, tokenizer: { - source: MOONSHINE_TOKENIZER, + source: MOONSHINE_TINY.tokenizerSource, bos: 1, // FIXME: this is a placeholder and needs to be changed eos: 2, // FIXME: this is a placeholder and needs to be changed }, @@ -43,11 +37,11 @@ const moonshineTinyModelConfig = { const whisperTinyMultilingualModelConfig = { sources: { - encoder: WHISPER_TINY_MULTILINGUAL_ENCODER, - decoder: WHISPER_TINY_MULTILINGUAL_DECODER, + encoder: WHISPER_TINY_MULTILINGUAL.encoderSource, + decoder: WHISPER_TINY_MULTILINGUAL.decoderSource, }, tokenizer: { - source: WHISPER_TINY_MULTILINGUAL_TOKENIZER, + source: WHISPER_TINY_MULTILINGUAL.tokenizerSource, bos: 50258, // FIXME: this is a placeholder and needs to be changed eos: 50257, // FIXME: this is a placeholder and needs to be changed }, diff --git a/packages/react-native-executorch/src/controllers/SpeechToTextController.ts b/packages/react-native-executorch/src/controllers/SpeechToTextController.ts index 3a53074057..7277979559 100644 --- a/packages/react-native-executorch/src/controllers/SpeechToTextController.ts +++ b/packages/react-native-executorch/src/controllers/SpeechToTextController.ts @@ -127,7 +127,7 @@ export class SpeechToTextController { // create a separate class for multilingual version of Whisper, since it is the same. We just need // the distinction here, in TS, for start tokens and such. If we introduce // more versions of Whisper, such as the small one, this should be refactored. - modelName = 'whisper'; + modelName = AvailableModels.WHISPER; } try { diff --git a/packages/react-native-executorch/src/types/stt.ts b/packages/react-native-executorch/src/types/stt.ts index 783316e857..a8074ff786 100644 --- a/packages/react-native-executorch/src/types/stt.ts +++ b/packages/react-native-executorch/src/types/stt.ts @@ -90,4 +90,8 @@ export enum SpeechToTextLanguage { Yiddish = 'yi', } -export type AvailableModels = 'whisper' | 'moonshine' | 'whisperMultilingual'; +export enum AvailableModels { + WHISPER = 'whisper', + MOONSHINE = 'moonshine', + WHISPER_MULTILINGUAL = 'whisperMultilingual', +} From 8a50b6b3407e2df9b7e095126b54aa1569fafdc8 Mon Sep 17 00:00:00 2001 From: jakmro Date: Thu, 31 Jul 2025 02:01:20 +0200 Subject: [PATCH 03/20] refactor: Change model constants to preset objects v2 --- .../app/classification/index.tsx | 2 +- .../app/image_segmentation/index.tsx | 2 +- .../app/object_detection/index.tsx | 2 +- apps/computer-vision/app/ocr/index.tsx | 2 +- .../app/ocr_vertical/index.tsx | 2 +- .../app/style_transfer/index.tsx | 2 +- apps/llm/app/llm/index.tsx | 2 +- apps/llm/app/llm_structured_output/index.tsx | 2 +- apps/llm/app/llm_tool_calling/index.tsx | 2 +- apps/llm/app/voice_chat/index.tsx | 4 +- .../screens/SpeechToTextScreen.tsx | 2 +- .../app/clip-embeddings/index.tsx | 4 +- .../app/text-embeddings/index.tsx | 2 +- .../src/constants/ocr/models.ts | 20 +++----- .../src/controllers/SpeechToTextController.ts | 10 ++-- .../computer_vision/useClassification.ts | 9 ++-- .../computer_vision/useImageEmbeddings.ts | 13 +++-- .../computer_vision/useImageSegmentation.ts | 9 ++-- .../src/hooks/computer_vision/useOCR.ts | 36 +++++++------ .../computer_vision/useObjectDetection.ts | 9 ++-- .../hooks/computer_vision/useStyleTransfer.ts | 6 +-- .../hooks/computer_vision/useVerticalOCR.ts | 51 +++++++++---------- .../src/hooks/general/useExecutorchModule.ts | 9 ++-- .../natural_language_processing/useLLM.ts | 51 +++++++++++-------- .../useSpeechToText.ts | 47 +++++++++-------- .../useTextEmbeddings.ts | 14 +++-- .../useTokenizer.ts | 12 ++--- .../src/hooks/useNonStaticModule.ts | 44 ++++++++-------- .../computer_vision/ClassificationModule.ts | 6 +-- .../computer_vision/ImageEmbeddingsModule.ts | 6 +-- .../ImageSegmentationModule.ts | 6 +-- .../src/modules/computer_vision/OCRModule.ts | 25 +++++---- .../computer_vision/ObjectDetectionModule.ts | 6 +-- .../computer_vision/StyleTransferModule.ts | 6 +-- .../computer_vision/VerticalOCRModule.ts | 31 ++++++----- .../src/modules/general/ExecutorchModule.ts | 6 +-- .../natural_language_processing/LLMModule.ts | 25 ++++----- .../SpeechToTextModule.ts | 30 +++++------ .../TextEmbeddingsModule.ts | 9 ++-- .../TokenizerModule.ts | 4 +- 40 files changed, 257 insertions(+), 273 deletions(-) diff --git a/apps/computer-vision/app/classification/index.tsx b/apps/computer-vision/app/classification/index.tsx index 19483d6996..8769ba9e3c 100644 --- a/apps/computer-vision/app/classification/index.tsx +++ b/apps/computer-vision/app/classification/index.tsx @@ -13,7 +13,7 @@ export default function ClassificationScreen() { ); const [imageUri, setImageUri] = useState(''); - const model = useClassification(EFFICIENTNET_V2_S); + const model = useClassification({ model: EFFICIENTNET_V2_S }); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/computer-vision/app/image_segmentation/index.tsx b/apps/computer-vision/app/image_segmentation/index.tsx index 29965d540a..8dfc7594e5 100644 --- a/apps/computer-vision/app/image_segmentation/index.tsx +++ b/apps/computer-vision/app/image_segmentation/index.tsx @@ -61,7 +61,7 @@ const numberToColor: number[][] = [ ]; export default function ImageSegmentationScreen() { - const model = useImageSegmentation(DEEPLAB_V3_RESNET50); + const model = useImageSegmentation({ model: DEEPLAB_V3_RESNET50 }); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/computer-vision/app/object_detection/index.tsx b/apps/computer-vision/app/object_detection/index.tsx index c0aed6a172..3ce52c409f 100644 --- a/apps/computer-vision/app/object_detection/index.tsx +++ b/apps/computer-vision/app/object_detection/index.tsx @@ -20,7 +20,7 @@ export default function ObjectDetectionScreen() { height: number; }>(); - const ssdLite = useObjectDetection(SSDLITE_320_MOBILENET_V3_LARGE); + const ssdLite = useObjectDetection({ model: SSDLITE_320_MOBILENET_V3_LARGE }); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(ssdLite.isGenerating); diff --git a/apps/computer-vision/app/ocr/index.tsx b/apps/computer-vision/app/ocr/index.tsx index 47dff1bbd5..b2ba8d04dc 100644 --- a/apps/computer-vision/app/ocr/index.tsx +++ b/apps/computer-vision/app/ocr/index.tsx @@ -16,7 +16,7 @@ export default function OCRScreen() { height: number; }>(); - const model = useOCR(OCR_ENGLISH); + const model = useOCR({ model: OCR_ENGLISH }); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/computer-vision/app/ocr_vertical/index.tsx b/apps/computer-vision/app/ocr_vertical/index.tsx index 6879b81563..040c709c63 100644 --- a/apps/computer-vision/app/ocr_vertical/index.tsx +++ b/apps/computer-vision/app/ocr_vertical/index.tsx @@ -16,7 +16,7 @@ export default function VerticalOCRScree() { height: number; }>(); const model = useVerticalOCR({ - ...VERTICAL_OCR_ENGLISH, + model: VERTICAL_OCR_ENGLISH, independentCharacters: true, }); const { setGlobalGenerating } = useContext(GeneratingContext); diff --git a/apps/computer-vision/app/style_transfer/index.tsx b/apps/computer-vision/app/style_transfer/index.tsx index 7cec152ac7..0075d1a532 100644 --- a/apps/computer-vision/app/style_transfer/index.tsx +++ b/apps/computer-vision/app/style_transfer/index.tsx @@ -11,7 +11,7 @@ import { GeneratingContext } from '../../context'; import ScreenWrapper from '../../ScreenWrapper'; export default function StyleTransferScreen() { - const model = useStyleTransfer(STYLE_TRANSFER_CANDY); + const model = useStyleTransfer({ model: STYLE_TRANSFER_CANDY }); const { setGlobalGenerating } = useContext(GeneratingContext); useEffect(() => { setGlobalGenerating(model.isGenerating); diff --git a/apps/llm/app/llm/index.tsx b/apps/llm/app/llm/index.tsx index b9eee28f2c..f49bab99fe 100644 --- a/apps/llm/app/llm/index.tsx +++ b/apps/llm/app/llm/index.tsx @@ -30,7 +30,7 @@ function LLMScreen() { const textInputRef = useRef(null); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM(LLAMA3_2_1B_QLORA); + const llm = useLLM({ model: LLAMA3_2_1B_QLORA }); useEffect(() => { if (llm.error) { diff --git a/apps/llm/app/llm_structured_output/index.tsx b/apps/llm/app/llm_structured_output/index.tsx index 9bdf6deb95..d850a61986 100644 --- a/apps/llm/app/llm_structured_output/index.tsx +++ b/apps/llm/app/llm_structured_output/index.tsx @@ -73,7 +73,7 @@ function LLMScreen() { const textInputRef = useRef(null); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM(QWEN3_1_7B_QUANTIZED); // try out 4B model if 1.7B struggles with following structured output + const llm = useLLM({ model: QWEN3_1_7B_QUANTIZED }); // try out 4B model if 1.7B struggles with following structured output useEffect(() => { setGlobalGenerating(llm.isGenerating); diff --git a/apps/llm/app/llm_tool_calling/index.tsx b/apps/llm/app/llm_tool_calling/index.tsx index 5e1440ceab..df177f7c67 100644 --- a/apps/llm/app/llm_tool_calling/index.tsx +++ b/apps/llm/app/llm_tool_calling/index.tsx @@ -39,7 +39,7 @@ function LLMToolCallingScreen() { const textInputRef = useRef(null); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM(HAMMER2_1_1_5B_QUANTIZED); + const llm = useLLM({ model: HAMMER2_1_1_5B_QUANTIZED }); useEffect(() => { setGlobalGenerating(llm.isGenerating); diff --git a/apps/llm/app/voice_chat/index.tsx b/apps/llm/app/voice_chat/index.tsx index 67a7294e00..bab7dfb115 100644 --- a/apps/llm/app/voice_chat/index.tsx +++ b/apps/llm/app/voice_chat/index.tsx @@ -67,9 +67,9 @@ function VoiceChatScreen() { const messageRecorded = useRef(false); const { setGlobalGenerating } = useContext(GeneratingContext); - const llm = useLLM(QWEN3_0_6B_QUANTIZED); + const llm = useLLM({ model: QWEN3_0_6B_QUANTIZED }); const speechToText = useSpeechToText({ - ...MOONSHINE_TINY, + model: MOONSHINE_TINY, windowSize: 3, overlapSeconds: 1.2, }); diff --git a/apps/speech-to-text/screens/SpeechToTextScreen.tsx b/apps/speech-to-text/screens/SpeechToTextScreen.tsx index bae8f0998a..0d9922e5dd 100644 --- a/apps/speech-to-text/screens/SpeechToTextScreen.tsx +++ b/apps/speech-to-text/screens/SpeechToTextScreen.tsx @@ -54,7 +54,7 @@ export const SpeechToTextScreen = () => { sequence, error, transcribe, - } = useSpeechToText(MOONSHINE_TINY); + } = useSpeechToText({ model: MOONSHINE_TINY }); const loadAudio = async (url: string) => { const audioContext = new AudioContext({ sampleRate: 16e3 }); diff --git a/apps/text-embeddings/app/clip-embeddings/index.tsx b/apps/text-embeddings/app/clip-embeddings/index.tsx index 924d1f1a13..7a53a77dfb 100644 --- a/apps/text-embeddings/app/clip-embeddings/index.tsx +++ b/apps/text-embeddings/app/clip-embeddings/index.tsx @@ -28,8 +28,8 @@ export default function ClipEmbeddingsScreenWrapper() { } function ClipEmbeddingsScreen() { - const textModel = useTextEmbeddings(CLIP_VIT_BASE_PATCH32_TEXT); - const imageModel = useImageEmbeddings(CLIP_VIT_BASE_PATCH32_IMAGE); + const textModel = useTextEmbeddings({ model: CLIP_VIT_BASE_PATCH32_TEXT }); + const imageModel = useImageEmbeddings({ model: CLIP_VIT_BASE_PATCH32_IMAGE }); const [inputSentence, setInputSentence] = useState(''); const [sentencesWithEmbeddings, setSentencesWithEmbeddings] = useState< diff --git a/apps/text-embeddings/app/text-embeddings/index.tsx b/apps/text-embeddings/app/text-embeddings/index.tsx index 9c6d9b6b93..3e43004dde 100644 --- a/apps/text-embeddings/app/text-embeddings/index.tsx +++ b/apps/text-embeddings/app/text-embeddings/index.tsx @@ -22,7 +22,7 @@ export default function TextEmbeddingsScreenWrapper() { } function TextEmbeddingsScreen() { - const model = useTextEmbeddings(ALL_MINILM_L6_V2); + const model = useTextEmbeddings({ model: ALL_MINILM_L6_V2 }); const [inputSentence, setInputSentence] = useState(''); const [sentencesWithEmbeddings, setSentencesWithEmbeddings] = useState< diff --git a/packages/react-native-executorch/src/constants/ocr/models.ts b/packages/react-native-executorch/src/constants/ocr/models.ts index a5bef1b181..16b7c7654c 100644 --- a/packages/react-native-executorch/src/constants/ocr/models.ts +++ b/packages/react-native-executorch/src/constants/ocr/models.ts @@ -101,11 +101,9 @@ const createOCRObject = ( ) => { return { detectorSource: DETECTOR_CRAFT_800_MODEL, - recognizerSources: { - recognizerLarge, - recognizerMedium, - recognizerSmall, - }, + recognizerLarge, + recognizerMedium, + recognizerSmall, language, }; }; @@ -116,14 +114,10 @@ const createVerticalOCRObject = ( language: keyof typeof symbols ) => { return { - detectorSources: { - detectorLarge: DETECTOR_CRAFT_1280_MODEL, - detectorNarrow: DETECTOR_CRAFT_320_MODEL, - }, - recognizerSources: { - recognizerLarge, - recognizerSmall, - }, + detectorLarge: DETECTOR_CRAFT_1280_MODEL, + detectorNarrow: DETECTOR_CRAFT_320_MODEL, + recognizerLarge, + recognizerSmall, language, }; }; diff --git a/packages/react-native-executorch/src/controllers/SpeechToTextController.ts b/packages/react-native-executorch/src/controllers/SpeechToTextController.ts index 7277979559..e773c17547 100644 --- a/packages/react-native-executorch/src/controllers/SpeechToTextController.ts +++ b/packages/react-native-executorch/src/controllers/SpeechToTextController.ts @@ -101,7 +101,7 @@ export class SpeechToTextController { try { const tokenizerLoadPromise = this.tokenizerModule.load( - tokenizerSource || this.config.tokenizer.source + { tokenizerSource: tokenizerSource || this.config.tokenizer.source } ); const pathsPromise = ResourceFetcher.fetch( onDownloadProgressCallback, @@ -414,7 +414,7 @@ export class SpeechToTextController { const chunk = this.waveform.slice( 0, this.windowSize + - this.overlapSeconds * (1 + Number(this.seqs.length > 0)) + this.overlapSeconds * (1 + Number(this.seqs.length > 0)) ); this.chunks = [chunk]; //save last chunk for STREAMING_ACTION.STOP this.waveform = this.waveform.slice( @@ -435,9 +435,9 @@ export class SpeechToTextController { // pad remaining waveform data with previous chunk to this.windowSize + 2 * this.overlapSeconds const chunk = this.chunks.length ? [ - ...this.chunks[0]!.slice(0, this.windowSize), - ...this.waveform, - ].slice(-this.windowSize - 2 * this.overlapSeconds) + ...this.chunks[0]!.slice(0, this.windowSize), + ...this.waveform, + ].slice(-this.windowSize - 2 * this.overlapSeconds) : this.waveform; this.waveform = []; diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useClassification.ts b/packages/react-native-executorch/src/hooks/computer_vision/useClassification.ts index 2c4dcec699..06088ec764 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useClassification.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useClassification.ts @@ -3,16 +3,13 @@ import { useNonStaticModule } from '../useNonStaticModule'; import { ClassificationModule } from '../../modules/computer_vision/ClassificationModule'; interface Props { - modelSource: ResourceSource; + model: { modelSource: ResourceSource }; preventLoad?: boolean; } -export const useClassification = ({ - modelSource, - preventLoad = false, -}: Props) => +export const useClassification = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: ClassificationModule, - loadArgs: [modelSource], + model, preventLoad: preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useImageEmbeddings.ts b/packages/react-native-executorch/src/hooks/computer_vision/useImageEmbeddings.ts index e81de1855c..2fff4d0412 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useImageEmbeddings.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useImageEmbeddings.ts @@ -2,15 +2,14 @@ import { ImageEmbeddingsModule } from '../../modules/computer_vision/ImageEmbedd import { ResourceSource } from '../../types/common'; import { useNonStaticModule } from '../useNonStaticModule'; -export const useImageEmbeddings = ({ - modelSource, - preventLoad = false, -}: { - modelSource: ResourceSource; +interface Props { + model: { modelSource: ResourceSource }; preventLoad?: boolean; -}) => +} + +export const useImageEmbeddings = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: ImageEmbeddingsModule, - loadArgs: [modelSource], + model, preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useImageSegmentation.ts b/packages/react-native-executorch/src/hooks/computer_vision/useImageSegmentation.ts index ff2ce941d2..31e26d059e 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useImageSegmentation.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useImageSegmentation.ts @@ -3,16 +3,13 @@ import { ImageSegmentationModule } from '../../modules/computer_vision/ImageSegm import { ResourceSource } from '../../types/common'; interface Props { - modelSource: ResourceSource; + model: { modelSource: ResourceSource }; preventLoad?: boolean; } -export const useImageSegmentation = ({ - modelSource, - preventLoad = false, -}: Props) => +export const useImageSegmentation = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: ImageSegmentationModule, - loadArgs: [modelSource], + model, preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts b/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts index e534bca28c..0b0020b054 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts @@ -12,18 +12,16 @@ interface OCRModule { } export const useOCR = ({ - detectorSource, - recognizerSources, - language = 'en', + model, preventLoad = false, }: { - detectorSource: ResourceSource; - recognizerSources: { + model: { + detectorSource: ResourceSource; recognizerLarge: ResourceSource; recognizerMedium: ResourceSource; recognizerSmall: ResourceSource; + language: OCRLanguage; }; - language?: OCRLanguage; preventLoad?: boolean; }): OCRModule => { const [error, setError] = useState(null); @@ -31,7 +29,7 @@ export const useOCR = ({ const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const model = useMemo( + const _model = useMemo( () => new OCRController({ modelDownloadProgressCallback: setDownloadProgress, @@ -44,19 +42,27 @@ export const useOCR = ({ useEffect(() => { const loadModel = async () => { - await model.loadModel(detectorSource, recognizerSources, language); + await _model.loadModel( + model.detectorSource, + { + recognizerLarge: model.recognizerLarge, + recognizerMedium: model.recognizerMedium, + recognizerSmall: model.recognizerSmall, + }, + model.language + ); }; if (!preventLoad) { loadModel(); } - // eslint-disable-next-line react-hooks/exhaustive-deps }, [ - model, - detectorSource, - language, - // eslint-disable-next-line react-hooks/exhaustive-deps - JSON.stringify(recognizerSources), + _model, + model.detectorSource, + model.recognizerLarge, + model.recognizerMedium, + model.recognizerSmall, + model.language, preventLoad, ]); @@ -64,7 +70,7 @@ export const useOCR = ({ error, isReady, isGenerating, - forward: model.forward, + forward: _model.forward, downloadProgress, }; }; diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useObjectDetection.ts b/packages/react-native-executorch/src/hooks/computer_vision/useObjectDetection.ts index 4d674f3e6b..181b770f4a 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useObjectDetection.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useObjectDetection.ts @@ -3,16 +3,13 @@ import { useNonStaticModule } from '../useNonStaticModule'; import { ObjectDetectionModule } from '../../modules/computer_vision/ObjectDetectionModule'; interface Props { - modelSource: ResourceSource; + model: { modelSource: ResourceSource }; preventLoad?: boolean; } -export const useObjectDetection = ({ - modelSource, - preventLoad = false, -}: Props) => +export const useObjectDetection = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: ObjectDetectionModule, - loadArgs: [modelSource], + model, preventLoad: preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useStyleTransfer.ts b/packages/react-native-executorch/src/hooks/computer_vision/useStyleTransfer.ts index 07a68f4a94..5e70dc9707 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useStyleTransfer.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useStyleTransfer.ts @@ -3,13 +3,13 @@ import { useNonStaticModule } from '../useNonStaticModule'; import { StyleTransferModule } from '../../modules/computer_vision/StyleTransferModule'; interface Props { - modelSource: ResourceSource; + model: { modelSource: ResourceSource }; preventLoad?: boolean; } -export const useStyleTransfer = ({ modelSource, preventLoad = false }: Props) => +export const useStyleTransfer = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: StyleTransferModule, - loadArgs: [modelSource], + model, preventLoad: preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts b/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts index 7f193c3719..cccdc58477 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts @@ -12,21 +12,17 @@ interface OCRModule { } export const useVerticalOCR = ({ - detectorSources, - recognizerSources, - language = 'en', + model, independentCharacters = false, preventLoad = false, }: { - detectorSources: { + model: { detectorLarge: ResourceSource; detectorNarrow: ResourceSource; - }; - recognizerSources: { recognizerLarge: ResourceSource; recognizerSmall: ResourceSource; + language: OCRLanguage; }; - language?: OCRLanguage; independentCharacters?: boolean; preventLoad?: boolean; }): OCRModule => { @@ -35,7 +31,7 @@ export const useVerticalOCR = ({ const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const model = useMemo( + const _model = useMemo( () => new VerticalOCRController({ modelDownloadProgressCallback: setDownloadProgress, @@ -47,27 +43,30 @@ export const useVerticalOCR = ({ ); useEffect(() => { - const loadModel = async () => { - await model.loadModel( - detectorSources, - recognizerSources, - language, + if (preventLoad) return; + + (async () => { + await _model.loadModel( + { + detectorLarge: model.detectorLarge, + detectorNarrow: model.detectorNarrow, + }, + { + recognizerLarge: model.recognizerLarge, + recognizerSmall: model.recognizerSmall, + }, + model.language, independentCharacters ); - }; - - if (!preventLoad) { - loadModel(); - } - // eslint-disable-next-line react-hooks/exhaustive-deps + })(); }, [ - model, - // eslint-disable-next-line react-hooks/exhaustive-deps - JSON.stringify(detectorSources), - language, + _model, + model.detectorLarge, + model.detectorNarrow, + model.recognizerLarge, + model.recognizerSmall, + model.language, independentCharacters, - // eslint-disable-next-line react-hooks/exhaustive-deps - JSON.stringify(recognizerSources), preventLoad, ]); @@ -75,7 +74,7 @@ export const useVerticalOCR = ({ error, isReady, isGenerating, - forward: model.forward, + forward: _model.forward, downloadProgress, }; }; diff --git a/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts b/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts index 42826a07a5..e5041c3d48 100644 --- a/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts +++ b/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts @@ -3,16 +3,13 @@ import { ResourceSource } from '../../types/common'; import { useNonStaticModule } from '../useNonStaticModule'; interface Props { - modelSource: ResourceSource; + model: { modelSource: ResourceSource }; preventLoad?: boolean; } -export const useExecutorchModule = ({ - modelSource, - preventLoad = false, -}: Props) => +export const useExecutorchModule = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: ExecutorchModule, - loadArgs: [modelSource], + model, preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts index c8522fa1e6..ee6d32f955 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts @@ -13,14 +13,14 @@ import { LLMController } from '../../controllers/LLMController'; Hook version of LLMModule */ export const useLLM = ({ - modelSource, - tokenizerSource, - tokenizerConfigSource, + model, preventLoad = false, }: { - modelSource: ResourceSource; - tokenizerSource: ResourceSource; - tokenizerConfigSource: ResourceSource; + model: { + modelSource: ResourceSource; + tokenizerSource: ResourceSource; + tokenizerConfigSource: ResourceSource; + }; preventLoad?: boolean; }): LLMType => { const [token, setToken] = useState(''); @@ -36,7 +36,7 @@ export const useLLM = ({ setResponse((prevResponse) => prevResponse + newToken); }, []); - const model = useMemo( + const _model = useMemo( () => new LLMController({ tokenCallback: tokenCallback, @@ -55,10 +55,10 @@ export const useLLM = ({ (async () => { try { - await model.load({ - modelSource, - tokenizerSource, - tokenizerConfigSource, + await _model.load({ + modelSource: model.modelSource, + tokenizerSource: model.tokenizerSource, + tokenizerConfigSource: model.tokenizerConfigSource, onDownloadProgressCallback: setDownloadProgress, }); } catch (e) { @@ -67,9 +67,15 @@ export const useLLM = ({ })(); return () => { - model.delete(); + _model.delete(); }; - }, [modelSource, tokenizerSource, tokenizerConfigSource, preventLoad, model]); + }, [ + _model, + model.modelSource, + model.tokenizerSource, + model.tokenizerConfigSource, + preventLoad, + ]); // memoization of returned functions const configure = useCallback( @@ -79,31 +85,32 @@ export const useLLM = ({ }: { chatConfig?: Partial; toolsConfig?: ToolsConfig; - }) => model.configure({ chatConfig, toolsConfig }), - [model] + }) => _model.configure({ chatConfig, toolsConfig }), + [_model] ); const generate = useCallback( (messages: Message[], tools?: LLMTool[]) => { setResponse(''); - return model.generate(messages, tools); + return _model.generate(messages, tools); }, - [model] + [_model] ); const sendMessage = useCallback( (message: string) => { setResponse(''); - return model.sendMessage(message); + return _model.sendMessage(message); }, - [model] + [_model] ); const deleteMessage = useCallback( - (index: number) => model.deleteMessage(index), - [model] + (index: number) => _model.deleteMessage(index), + [_model] ); - const interrupt = useCallback(() => model.interrupt(), [model]); + + const interrupt = useCallback(() => _model.interrupt(), [_model]); return { messageHistory, diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts index e22b055b91..1bb493e105 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts @@ -23,19 +23,18 @@ interface SpeechToTextModule { } export const useSpeechToText = ({ - modelName, - encoderSource, - decoderSource, - tokenizerSource, + model, overlapSeconds, windowSize, streamingConfig, preventLoad = false, }: { - modelName: AvailableModels; - encoderSource?: ResourceSource; - decoderSource?: ResourceSource; - tokenizerSource?: ResourceSource; + model: { + modelName: AvailableModels; + encoderSource: ResourceSource; + decoderSource: ResourceSource; + tokenizerSource: ResourceSource; + }; overlapSeconds?: ConstructorParameters< typeof SpeechToTextController >['0']['overlapSeconds']; @@ -53,7 +52,7 @@ export const useSpeechToText = ({ const [isGenerating, setIsGenerating] = useState(false); const [error, setError] = useState(); - const model = useMemo( + const _model = useMemo( () => new SpeechToTextController({ transcribeCallback: setSequence, @@ -65,16 +64,16 @@ export const useSpeechToText = ({ ); useEffect(() => { - model.configureStreaming(overlapSeconds, windowSize, streamingConfig); - }, [model, overlapSeconds, windowSize, streamingConfig]); + _model.configureStreaming(overlapSeconds, windowSize, streamingConfig); + }, [_model, overlapSeconds, windowSize, streamingConfig]); useEffect(() => { const loadModel = async () => { - await model.load({ - modelName, - encoderSource, - decoderSource, - tokenizerSource, + await _model.load({ + modelName: model.modelName, + encoderSource: model.encoderSource, + decoderSource: model.decoderSource, + tokenizerSource: model.tokenizerSource, onDownloadProgressCallback: setDownloadProgress, }); }; @@ -82,11 +81,11 @@ export const useSpeechToText = ({ loadModel(); } }, [ - model, - modelName, - encoderSource, - decoderSource, - tokenizerSource, + _model, + model.modelName, + model.encoderSource, + model.decoderSource, + model.tokenizerSource, preventLoad, ]); @@ -94,15 +93,15 @@ export const useSpeechToText = ({ isReady, isGenerating, downloadProgress, - configureStreaming: model.configureStreaming, + configureStreaming: _model.configureStreaming, sequence, error, transcribe: (waveform: number[], audioLanguage?: SpeechToTextLanguage) => - model.transcribe(waveform, audioLanguage), + _model.transcribe(waveform, audioLanguage), streamingTranscribe: ( streamAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage - ) => model.streamingTranscribe(streamAction, waveform, audioLanguage), + ) => _model.streamingTranscribe(streamAction, waveform, audioLanguage), }; }; diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useTextEmbeddings.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useTextEmbeddings.ts index 058ed77eb9..18a7e1e453 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useTextEmbeddings.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useTextEmbeddings.ts @@ -3,18 +3,16 @@ import { ResourceSource } from '../../types/common'; import { useNonStaticModule } from '../useNonStaticModule'; interface Props { - modelSource: ResourceSource; - tokenizerSource: ResourceSource; + model: { + modelSource: ResourceSource; + tokenizerSource: ResourceSource; + }; preventLoad?: boolean; } -export const useTextEmbeddings = ({ - modelSource, - tokenizerSource, - preventLoad = false, -}: Props) => +export const useTextEmbeddings = ({ model, preventLoad = false }: Props) => useNonStaticModule({ module: TextEmbeddingsModule, - loadArgs: [modelSource, tokenizerSource], + model, preventLoad, }); diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts index 9748d0fafe..5101bbb8b4 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts @@ -4,17 +4,17 @@ import { ResourceSource } from '../../types/common'; import { ETError, getError } from '../../Error'; export const useTokenizer = ({ - tokenizerSource, + tokenizer, preventLoad = false, }: { - tokenizerSource: ResourceSource; + tokenizer: { tokenizerSource: ResourceSource }; preventLoad?: boolean; }) => { const [error, setError] = useState(null); const [isReady, setIsReady] = useState(false); const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const model = useMemo(() => new TokenizerModule(), []); + const _tokenizer = useMemo(() => new TokenizerModule(), []); useEffect(() => { if (preventLoad) return; @@ -23,13 +23,13 @@ export const useTokenizer = ({ setError(null); try { setIsReady(false); - await model.load(tokenizerSource, setDownloadProgress); + await _tokenizer.load({ tokenizerSource: tokenizer.tokenizerSource }, setDownloadProgress); setIsReady(true); } catch (err) { setError((err as Error).message); } })(); - }, [model, tokenizerSource, preventLoad]); + }, [_tokenizer, tokenizer.tokenizerSource, preventLoad]); const stateWrapper = Promise>(fn: T) => { return (...args: Parameters): Promise> => { @@ -37,7 +37,7 @@ export const useTokenizer = ({ if (isGenerating) throw new Error(getError(ETError.ModelGenerating)); try { setIsGenerating(true); - return fn.apply(model, args); + return fn.apply(_tokenizer, args); } finally { setIsGenerating(false); } diff --git a/packages/react-native-executorch/src/hooks/useNonStaticModule.ts b/packages/react-native-executorch/src/hooks/useNonStaticModule.ts index 04a1c8e2d9..05c15b347d 100644 --- a/packages/react-native-executorch/src/hooks/useNonStaticModule.ts +++ b/packages/react-native-executorch/src/hooks/useNonStaticModule.ts @@ -18,47 +18,47 @@ export const useNonStaticModule = < ForwardReturn extends Awaited>, >({ module, - loadArgs, + model, preventLoad = false, }: { module: ModuleConstructor; - loadArgs: LoadArgs; + model: LoadArgs[0]; preventLoad?: boolean; }) => { const [error, setError] = useState(null); const [isReady, setIsReady] = useState(false); const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const model = useMemo(() => new module(), [module]); + const _model = useMemo(() => new module(), [module]); useEffect(() => { - if (!preventLoad) { - (async () => { - setDownloadProgress(0); - setError(null); - try { - setIsReady(false); - await model.load(...loadArgs, setDownloadProgress); - setIsReady(true); - } catch (err) { - setError((err as Error).message); - } - })(); + if (preventLoad) return; + + (async () => { + setDownloadProgress(0); + setError(null); + try { + setIsReady(false); + await _model.load(model, setDownloadProgress); + setIsReady(true); + } catch (err) { + setError((err as Error).message); + } + })(); + + return () => { + _model.delete(); + }; - return () => { - model.delete(); - }; - } - return () => {}; // eslint-disable-next-line react-hooks/exhaustive-deps - }, [...loadArgs, preventLoad]); + }, [...Object.values(model), preventLoad]); const forward = async (...input: ForwardArgs): Promise => { if (!isReady) throw new Error(getError(ETError.ModuleNotLoaded)); if (isGenerating) throw new Error(getError(ETError.ModelGenerating)); try { setIsGenerating(true); - return await model.forward(...input); + return await _model.forward(...input); } finally { setIsGenerating(false); } diff --git a/packages/react-native-executorch/src/modules/computer_vision/ClassificationModule.ts b/packages/react-native-executorch/src/modules/computer_vision/ClassificationModule.ts index edbb36ba95..c7e23b3e52 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/ClassificationModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/ClassificationModule.ts @@ -5,12 +5,12 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class ClassificationModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); diff --git a/packages/react-native-executorch/src/modules/computer_vision/ImageEmbeddingsModule.ts b/packages/react-native-executorch/src/modules/computer_vision/ImageEmbeddingsModule.ts index 3ff8b2a5a0..133f31707d 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/ImageEmbeddingsModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/ImageEmbeddingsModule.ts @@ -5,12 +5,12 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class ImageEmbeddingsModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); diff --git a/packages/react-native-executorch/src/modules/computer_vision/ImageSegmentationModule.ts b/packages/react-native-executorch/src/modules/computer_vision/ImageSegmentationModule.ts index 733219fed4..a7d2b5d93c 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/ImageSegmentationModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/ImageSegmentationModule.ts @@ -6,12 +6,12 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class ImageSegmentationModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); diff --git a/packages/react-native-executorch/src/modules/computer_vision/OCRModule.ts b/packages/react-native-executorch/src/modules/computer_vision/OCRModule.ts index c7a28ef622..e7ffd4cd9a 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/OCRModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/OCRModule.ts @@ -5,29 +5,32 @@ import { OCRLanguage } from '../../types/ocr'; export class OCRModule { static module: OCRController; - static onDownloadProgressCallback = (_downloadProgress: number) => {}; - static async load( - detectorSource: ResourceSource, - recognizerSources: { + model: { + detectorSource: ResourceSource; recognizerLarge: ResourceSource; recognizerMedium: ResourceSource; recognizerSmall: ResourceSource; + language: OCRLanguage; }, - language: OCRLanguage = 'en' + onDownloadProgressCallback: (progress: number) => void = () => {} ) { this.module = new OCRController({ - modelDownloadProgressCallback: this.onDownloadProgressCallback, + modelDownloadProgressCallback: onDownloadProgressCallback, }); - await this.module.loadModel(detectorSource, recognizerSources, language); + await this.module.loadModel( + model.detectorSource, + { + recognizerLarge: model.recognizerLarge, + recognizerMedium: model.recognizerMedium, + recognizerSmall: model.recognizerSmall, + }, + model.language + ); } static async forward(input: string) { return await this.module.forward(input); } - - static onDownloadProgress(callback: (downloadProgress: number) => void) { - this.onDownloadProgressCallback = callback; - } } diff --git a/packages/react-native-executorch/src/modules/computer_vision/ObjectDetectionModule.ts b/packages/react-native-executorch/src/modules/computer_vision/ObjectDetectionModule.ts index 83e66b4da6..1fc16c4924 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/ObjectDetectionModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/ObjectDetectionModule.ts @@ -6,12 +6,12 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class ObjectDetectionModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); diff --git a/packages/react-native-executorch/src/modules/computer_vision/StyleTransferModule.ts b/packages/react-native-executorch/src/modules/computer_vision/StyleTransferModule.ts index 4a8ae94965..2969194a43 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/StyleTransferModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/StyleTransferModule.ts @@ -5,12 +5,12 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class StyleTransferModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); diff --git a/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts b/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts index 4c8b1120de..81ce4104aa 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts @@ -5,37 +5,36 @@ import { OCRLanguage } from '../../types/ocr'; export class VerticalOCRModule { static module: VerticalOCRController; - static onDownloadProgressCallback = (_downloadProgress: number) => {}; - static async load( - detectorSources: { + model: { detectorLarge: ResourceSource; detectorNarrow: ResourceSource; - }, - recognizerSources: { recognizerLarge: ResourceSource; recognizerSmall: ResourceSource; + language: OCRLanguage; + independentCharacters: boolean; }, - language: OCRLanguage = 'en', - independentCharacters: boolean = false + onDownloadProgressCallback: (progress: number) => void = () => {} ) { this.module = new VerticalOCRController({ - modelDownloadProgressCallback: this.onDownloadProgressCallback, + modelDownloadProgressCallback: onDownloadProgressCallback, }); await this.module.loadModel( - detectorSources, - recognizerSources, - language, - independentCharacters + { + detectorLarge: model.detectorLarge, + detectorNarrow: model.detectorNarrow, + }, + { + recognizerLarge: model.recognizerLarge, + recognizerSmall: model.recognizerSmall, + }, + model.language, + model.independentCharacters ); } static async forward(input: string) { return await this.module.forward(input); } - - static onDownloadProgress(callback: (downloadProgress: number) => void) { - this.onDownloadProgressCallback = callback; - } } diff --git a/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts b/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts index 8c8e7ec0a7..ba5f58b7f6 100644 --- a/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts +++ b/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts @@ -5,12 +5,12 @@ import { ResourceFetcher } from '../../utils/ResourceFetcher'; export class ExecutorchModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts index 520f07591d..8c5ebdbdb7 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts @@ -21,21 +21,18 @@ export class LLMModule { }); } - async load({ - modelSource, - tokenizerSource, - tokenizerConfigSource, - onDownloadProgressCallback, - }: { - modelSource: ResourceSource; - tokenizerSource: ResourceSource; - tokenizerConfigSource: ResourceSource; - onDownloadProgressCallback?: (_downloadProgress: number) => void; - }) { + async load( + model: { + modelSource: ResourceSource; + tokenizerSource: ResourceSource; + tokenizerConfigSource: ResourceSource; + }, + onDownloadProgressCallback: (progress: number) => void = () => {} + ) { await this.controller.load({ - modelSource, - tokenizerSource, - tokenizerConfigSource, + modelSource: model.modelSource, + tokenizerSource: model.tokenizerSource, + tokenizerConfigSource: model.tokenizerConfigSource, onDownloadProgressCallback, }); } diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/SpeechToTextModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/SpeechToTextModule.ts index 4f5ebf9a08..6bda902e38 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/SpeechToTextModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/SpeechToTextModule.ts @@ -31,24 +31,20 @@ export class SpeechToTextModule { }); } - async load({ - modelName, - encoderSource, - decoderSource, - tokenizerSource, - onDownloadProgressCallback, - }: { - modelName: AvailableModels; - encoderSource?: ResourceSource; - decoderSource?: ResourceSource; - tokenizerSource?: ResourceSource; - onDownloadProgressCallback?: (downloadProgress: number) => void; - }) { + async load( + model: { + modelName: AvailableModels; + encoderSource?: ResourceSource; + decoderSource?: ResourceSource; + tokenizerSource?: ResourceSource; + }, + onDownloadProgressCallback: (progress: number) => void = () => {} + ) { await this.module.load({ - modelName, - encoderSource, - decoderSource, - tokenizerSource, + modelName: model.modelName, + encoderSource: model.encoderSource, + decoderSource: model.decoderSource, + tokenizerSource: model.tokenizerSource, onDownloadProgressCallback, }); } diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts index c883de3ed5..ac73e27404 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts @@ -4,15 +4,14 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class TextEmbeddingsModule extends BaseNonStaticModule { async load( - modelSource: ResourceSource, - tokenizerSource: ResourceSource, - onDownloadProgressCallback: (_: number) => void = () => {} + model: { modelSource: ResourceSource; tokenizerSource: ResourceSource }, + onDownloadProgressCallback: (progress: number) => void = () => { } ): Promise { const modelPromise = ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + model.modelSource ); - const tokenizerPromise = ResourceFetcher.fetch(undefined, tokenizerSource); + const tokenizerPromise = ResourceFetcher.fetch(undefined, model.tokenizerSource); const [modelResult, tokenizerResult] = await Promise.all([ modelPromise, tokenizerPromise, diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts index 87e1965a37..df5d7460e9 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts @@ -5,12 +5,12 @@ export class TokenizerModule { nativeModule: any; async load( - modelSource: ResourceSource, + tokenizer: { tokenizerSource: ResourceSource }, onDownloadProgressCallback: (_: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - modelSource + tokenizer.tokenizerSource ); const path = paths?.[0]; if (!path) { From e99ef6a856c8fd7ecda6dcd0d86684a2b48737b7 Mon Sep 17 00:00:00 2001 From: jakmro Date: Thu, 31 Jul 2025 11:14:54 +0200 Subject: [PATCH 04/20] refactor: remove nesting in ExecutorchModule and its hook, extract independentCharacters from model in VerticalOCRModule --- .../src/hooks/general/useExecutorchModule.ts | 9 ++++++--- .../src/modules/computer_vision/VerticalOCRModule.ts | 4 ++-- .../src/modules/general/ExecutorchModule.ts | 4 ++-- 3 files changed, 10 insertions(+), 7 deletions(-) diff --git a/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts b/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts index e5041c3d48..72cd1999f5 100644 --- a/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts +++ b/packages/react-native-executorch/src/hooks/general/useExecutorchModule.ts @@ -3,13 +3,16 @@ import { ResourceSource } from '../../types/common'; import { useNonStaticModule } from '../useNonStaticModule'; interface Props { - model: { modelSource: ResourceSource }; + modelSource: ResourceSource; preventLoad?: boolean; } -export const useExecutorchModule = ({ model, preventLoad = false }: Props) => +export const useExecutorchModule = ({ + modelSource, + preventLoad = false, +}: Props) => useNonStaticModule({ module: ExecutorchModule, - model, + model: { modelSource }, preventLoad, }); diff --git a/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts b/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts index 81ce4104aa..aaa4e83a10 100644 --- a/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts +++ b/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts @@ -12,8 +12,8 @@ export class VerticalOCRModule { recognizerLarge: ResourceSource; recognizerSmall: ResourceSource; language: OCRLanguage; - independentCharacters: boolean; }, + independentCharacters: boolean, onDownloadProgressCallback: (progress: number) => void = () => {} ) { this.module = new VerticalOCRController({ @@ -30,7 +30,7 @@ export class VerticalOCRModule { recognizerSmall: model.recognizerSmall, }, model.language, - model.independentCharacters + independentCharacters ); } diff --git a/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts b/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts index ba5f58b7f6..6ae6337538 100644 --- a/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts +++ b/packages/react-native-executorch/src/modules/general/ExecutorchModule.ts @@ -5,12 +5,12 @@ import { ResourceFetcher } from '../../utils/ResourceFetcher'; export class ExecutorchModule extends BaseNonStaticModule { async load( - model: { modelSource: ResourceSource }, + modelSource: ResourceSource, onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, - model.modelSource + modelSource ); if (paths === null || paths.length < 1) { throw new Error('Download interrupted.'); From 33cdfe72196647492a5042efe694b979b62c6dd5 Mon Sep 17 00:00:00 2001 From: jakmro Date: Thu, 31 Jul 2025 13:10:25 +0200 Subject: [PATCH 05/20] refactor: update parameter type for onDownloadProgressCallback in TokenizerModule --- .../src/modules/natural_language_processing/TokenizerModule.ts | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts index df5d7460e9..47f3dd2f91 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/TokenizerModule.ts @@ -6,7 +6,7 @@ export class TokenizerModule { async load( tokenizer: { tokenizerSource: ResourceSource }, - onDownloadProgressCallback: (_: number) => void = () => {} + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const paths = await ResourceFetcher.fetch( onDownloadProgressCallback, From 6520c58aa9f37d5ecb552a0ff693fafcf2a36c87 Mon Sep 17 00:00:00 2001 From: jakmro Date: Thu, 31 Jul 2025 13:20:15 +0200 Subject: [PATCH 06/20] docs: update model loading syntax to use object parameters --- .../01-natural-language-processing/useLLM.md | 36 +++++++------ .../useSpeechToText.md | 31 +++++------ .../useTextEmbeddings.md | 12 ++--- .../useTokenizer.md | 8 +-- .../02-computer-vision/useClassification.md | 9 ++-- .../02-computer-vision/useImageEmbeddings.md | 7 +-- .../useImageSegmentation.md | 9 ++-- .../02-hooks/02-computer-vision/useOCR.md | 20 ++++--- .../02-computer-vision/useObjectDetection.md | 10 ++-- .../02-computer-vision/useStyleTransfer.md | 9 ++-- .../02-computer-vision/useVerticalOCR.md | 23 ++++---- .../LLMModule.md | 52 +++++++----------- .../SpeechToTextModule.md | 30 ++++++----- .../TextEmbeddingsModule.md | 23 +++++--- .../TokenizerModule.md | 23 ++++---- .../ClassificationModule.md | 18 ++++--- .../ImageEmbeddingsModule.md | 18 ++++--- .../ImageSegmentationModule.md | 10 +++- .../02-computer-vision/OCRModule.md | 54 ++++++------------- .../ObjectDetectionModule.md | 18 ++++--- .../02-computer-vision/StyleTransferModule.md | 18 ++++--- .../02-computer-vision/VerticalOCRModule.md | 53 +++++------------- .../ExecutorchModule.md | 14 ++--- 23 files changed, 241 insertions(+), 264 deletions(-) diff --git a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md index 927b31aba8..f5d16613a3 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md @@ -41,7 +41,7 @@ In order to load a model into the app, you need to run the following code: ```typescript import { useLLM, LLAMA3_2_1B } from 'react-native-executorch'; -const llm = useLLM(LLAMA3_2_1B); +const llm = useLLM({ model: LLAMA3_2_1B }); ```
@@ -50,11 +50,13 @@ The code snippet above fetches the model from the specified URL, loads it into m ### Arguments -**`modelSource`** - `ResourceSource` that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +**`model`** - Object containing the model source, tokenizer source, and tokenizer config source. -**`tokenizerSource`** - `ResourceSource` pointing to the JSON file which contains the tokenizer. +- **`modelSource`** - `ResourceSource` that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. -**`tokenizerConfigSource`** - `ResourceSource` pointing to the JSON file which contains the tokenizer config. +- **`tokenizerSource`** - `ResourceSource` pointing to the JSON file which contains the tokenizer. + +- **`tokenizerConfigSource`** - `ResourceSource` pointing to the JSON file which contains the tokenizer config. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -80,18 +82,18 @@ The code snippet above fetches the model from the specified URL, loads it into m ```typescript const useLLM: ({ - modelSource, - tokenizerSource, - tokenizerConfigSource, - preventLoad = false, + model, + preventLoad, }: { - modelSource: ResourceSource; - tokenizerSource: ResourceSource; - tokenizerConfigSource: ResourceSource; + model: { + modelSource: ResourceSource; + tokenizerSource: ResourceSource; + tokenizerConfigSource: ResourceSource; + }; preventLoad?: boolean; }) => LLMType; -interface LLMType { +export interface LLMType { messageHistory: Message[]; response: string; token: string; @@ -158,7 +160,7 @@ You can use functions returned from this hooks in two manners: To perform chat completion you can use the `generate` function. There is no return value. Instead, the `response` value is updated with each token. ```tsx -const llm = useLLM(LLAMA3_2_1B); +const llm = useLLM({ model: LLAMA3_2_1B }); const handleGenerate = () => { const chat = [ @@ -213,7 +215,7 @@ const TOOL_DEFINITIONS: LLMTool[] = [ }, ]; -const llm = useLLM(HAMMER2_1_1_5B); +const llm = useLLM({ model: HAMMER2_1_1_5B }); const handleGenerate = () => { const chat = [ @@ -272,7 +274,7 @@ To configure model (i.e. change system prompt, load initial conversation history In order to send a message to the model, one can use the following code: ```tsx -const llm = useLLM(LLAMA3_2_1B); +const llm = useLLM({ model: LLAMA3_2_1B }); const send = () => { const message = 'Hi, who are you?'; @@ -317,7 +319,7 @@ const TOOL_DEFINITIONS: LLMTool[] = [ }, ]; -const llm = useLLM(HAMMER2_1_1_5B); +const llm = useLLM({ model: HAMMER2_1_1_5B }); useEffect(() => { llm.configure({ @@ -393,7 +395,7 @@ const responseSchemaWithZod = z.object({ currency: z.optional(z.string().meta({ description: 'Currency of offer.' })), }); -const llm = useLLM(QWEN3_4B_QUANTIZED); +const llm = useLLM({ model: QWEN3_4B_QUANTIZED }); useEffect(() => { const formattingInstructions = getStructuredOutputPrompt(responseSchema); diff --git a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md index 710399c32a..5bd06bef20 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md @@ -37,7 +37,7 @@ import { useSpeechToText, MOONSHINE_TINY } from 'react-native-executorch'; import { AudioContext } from 'react-native-audio-api'; import * as FileSystem from 'expo-file-system'; -const { transcribe, error } = useSpeechToText(MOONSHINE_TINY); +const { transcribe, error } = useSpeechToText({ model: MOONSHINE_TINY }); const loadAudio = async (url: string) => { const audioContext = new AudioContext({ sampleRate: 16e3 }); @@ -66,26 +66,21 @@ Given that STT models can process audio no longer than 30 seconds, there is a ne ### Arguments -**`modelName`** -A literal of `"moonshine" | "whisper" | "whisperMultilingual` which serves as an identifier for which model should be used. +**`model`** - Object containing the model name, encoder source, decoder source, and tokenizer source. -**`encoderSource?`** -A string that specifies the location of a .pte file for the encoder. For further information on passing model sources, check out [Loading Models](../../01-fundamentals/02-loading-models.md). Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. +- **`modelName`** - A literal of `"moonshine" | "whisper" | "whisperMultilingual` which serves as an identifier for which model should be used. -**`decoderSource?`** -Analogous to the encoderSource, this takes in a string which is a source for the decoder part of the model. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. +- **`encoderSource?`** - A string that specifies the location of a .pte file for the encoder. For further information on passing model sources, check out [Loading Models](../../01-fundamentals/02-loading-models.md). Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. -**`tokenizerSource?`** -A string that specifies the location to the tokenizer for the model. This works just as the encoder and decoder do. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. +- **`decoderSource?`** - Analogous to the encoderSource, this takes in a string which is a source for the decoder part of the model. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. -**`overlapSeconds?`** -Specifies the length of overlap between consecutive audio chunks (expressed in seconds). Overrides `streamingConfig` argument. +- **`tokenizerSource?`** - A string that specifies the location to the tokenizer for the model. This works just as the encoder and decoder do. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. -**`windowSize?`** -Specifies the size of each audio chunk (expressed in seconds). Overrides `streamingConfig` argument. +**`overlapSeconds?`** - Specifies the length of overlap between consecutive audio chunks (expressed in seconds). Overrides `streamingConfig` argument. -**`streamingConfig?`** -Specifies config for both `overlapSeconds` and `windowSize` values. Three options are available: `fast`, `balanced` and `quality`. We discourage using `fast` config with `Whisper` model which while has the lowest latency to first token has the slowest overall speed. +**`windowSize?`** - Specifies the size of each audio chunk (expressed in seconds). Overrides `streamingConfig` argument. + +**`streamingConfig?`** - Specifies config for both `overlapSeconds` and `windowSize` values. Three options are available: `fast`, `balanced` and `quality`. We discourage using `fast` config with `Whisper` model which while has the lowest latency to first token has the slowest overall speed. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -219,7 +214,9 @@ import * as FileSystem from 'expo-file-system'; import { AudioContext } from 'react-native-audio-api'; function App() { - const { transcribe, sequence, error } = useSpeechToText(WHISPER_TINY); + const { transcribe, sequence, error } = useSpeechToText({ + model: WHISPER_TINY, + }); const loadAudio = async (url: string) => { const audioContext = new AudioContext({ sampleRate: 16e3 }); @@ -291,7 +288,7 @@ const float32ArrayFromPCMBinaryBuffer = (b64EncodedBuffer: string) => { function App() { const [isRecording, setIsRecording] = useState(false); const speechToText = useSpeechToText({ - ...MOONSHINE_TINY, + model: MOONSHINE_TINY, windowSize: 3, overlapSeconds: 1.2, }); diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md index 805ad9341f..768d4e4d09 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md @@ -26,7 +26,7 @@ It is recommended to use models provided by us, which are available at our [Hugg ```typescript import { useTextEmbeddings, ALL_MINILM_L6_V2 } from 'react-native-executorch'; -const model = useTextEmbeddings(ALL_MINILM_L6_V2); +const model = useTextEmbeddings({ model: ALL_MINILM_L6_V2 }); try { const embedding = await model.forward('Hello World!'); @@ -37,11 +37,11 @@ try { ### Arguments -**`modelSource`** -A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +**`model`** - Object containing the model source and tokenizer source. -**`tokenizerSource`** -A string that specifies the location of the tokenizer JSON file. +- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + +- **`tokenizerSource`** - A string that specifies the location of the tokenizer JSON file. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -75,7 +75,7 @@ const cosineSimilarity = (a: number[], b: number[]) => { }; function App() { - const model = useTextEmbeddings(ALL_MINILM_L6_V2); + const model = useTextEmbeddings({ model: ALL_MINILM_L6_V2 }); // ... diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md index 74fb8e54ea..eeb5bf3795 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md @@ -27,7 +27,7 @@ We are using [Hugging Face Tokenizers](https://huggingface.co/docs/tokenizers/in ```typescript import { useTokenizer, ALL_MINILM_L6_V2 } from 'react-native-executorch'; -const tokenizer = useTokenizer(ALL_MINILM_L6_V2); +const tokenizer = useTokenizer({ tokenizer: ALL_MINILM_L6_V2 }); const text = 'Hello, world!'; @@ -46,7 +46,9 @@ try { ## Arguments -**`tokenizerSource`** - A string that specifies the path or URI of the tokenizer JSON file. +**`tokenizer`** - Object containing the tokenizer source. + +- **`tokenizerSource`** - A string that specifies the location of the tokenizer JSON file. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -70,7 +72,7 @@ try { import { useTokenizer, ALL_MINILM_L6_V2 } from 'react-native-executorch'; function App() { - const tokenizer = useTokenizer(ALL_MINILM_L6_V2); + const tokenizer = useTokenizer({ tokenizer: ALL_MINILM_L6_V2 }); // ... diff --git a/docs/docs/02-hooks/02-computer-vision/useClassification.md b/docs/docs/02-hooks/02-computer-vision/useClassification.md index be702e3e6f..819aa0f7f0 100644 --- a/docs/docs/02-hooks/02-computer-vision/useClassification.md +++ b/docs/docs/02-hooks/02-computer-vision/useClassification.md @@ -17,7 +17,7 @@ It is recommended to use models provided by us, which are available at our [Hugg ```typescript import { useClassification, EFFICIENTNET_V2_S } from 'react-native-executorch'; -const model = useClassification(EFFICIENTNET_V2_S); +const model = useClassification({ model: EFFICIENTNET_V2_S }); const imageUri = 'file::///Users/.../cute_puppy.png'; @@ -30,8 +30,9 @@ try { ### Arguments -**`modelSource`** -A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -59,7 +60,7 @@ Images from external sources are stored in your application's temporary director import { useClassification, EFFICIENTNET_V2_S } from 'react-native-executorch'; function App() { - const model = useClassification(EFFICIENTNET_V2_S); + const model = useClassification({ model: EFFICIENTNET_V2_S }); // ... const imageUri = 'file:///Users/.../cute_puppy.png'; diff --git a/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md b/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md index f5758d60e3..8fa66d6802 100644 --- a/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md +++ b/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md @@ -30,7 +30,7 @@ import { CLIP_VIT_BASE_PATCH32_IMAGE, } from 'react-native-executorch'; -const model = useImageEmbeddings(CLIP_VIT_BASE_PATCH32_IMAGE); +const model = useImageEmbeddings({ model: CLIP_VIT_BASE_PATCH32_IMAGE }); try { const imageEmbedding = await model.forward('https://url-to-image.jpg'); @@ -41,8 +41,9 @@ try { ### Arguments -**`modelSource`** -A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. diff --git a/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md b/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md index e2b1c1a11b..031b0ea028 100644 --- a/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md +++ b/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md @@ -16,7 +16,7 @@ import { DEEPLAB_V3_RESNET50, } from 'react-native-executorch'; -const model = useImageSegmentation(DEEPLAB_V3_RESNET50); +const model = useImageSegmentation({ model: DEEPLAB_V3_RESNET50 }); const imageUri = 'file::///Users/.../cute_cat.png'; @@ -29,8 +29,9 @@ try { ### Arguments -**`modelSource`** -A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -65,7 +66,7 @@ Setting `resize` to true will make `forward` slower. ```typescript function App() { - const model = useImageSegmentation(DEEPLAB_V3_RESNET50); + const model = useImageSegmentation({ model: DEEPLAB_V3_RESNET50 }); // ... const imageUri = 'file::///Users/.../cute_cat.png'; diff --git a/docs/docs/02-hooks/02-computer-vision/useOCR.md b/docs/docs/02-hooks/02-computer-vision/useOCR.md index 1e757b281f..62500c26ee 100644 --- a/docs/docs/02-hooks/02-computer-vision/useOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useOCR.md @@ -14,7 +14,7 @@ It is recommended to use models provided by us, which are available at our [Hugg import { useOCR, OCR_ENGLISH } from 'react-native-executorch'; function App() { - const model = useOCR(OCR_ENGLISH); + const model = useOCR({ model: OCR_ENGLISH }); // ... for (const ocrDetection of await model.forward('https://url-to-image.jpg')) { @@ -118,20 +118,18 @@ interface OCRDetection { ### Arguments -**`detectorSource`** - A string that specifies the location of the detector binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +**`model`** - Object containing the model source. -**`recognizerSources`** - An object that specifies locations of the recognizers binary files. Each recognizer is composed of three models tailored to process images of varying widths. +- **`detectorSource`** - A string that specifies the location of the detector binary. +- **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. +- **`recognizerMedium`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 256 pixels. +- **`recognizerSmall`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels. +- **`language`** - A parameter that specifies the language of the text to be recognized by the OCR. -- `recognizerLarge` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. -- `recognizerMedium` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 256 pixels. -- `recognizerSmall` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels. +**`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. -**`language`** - A parameter that specifies the language of the text to be recognized by the OCR. - -**`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. - ### Returns The hook returns an object with the following properties: @@ -174,7 +172,7 @@ The `text` property contains the text recognized within detected text region. Th import { useOCR, OCR_ENGLISH } from 'react-native-executorch'; function App() { - const model = useOCR(OCR_ENGLISH); + const model = useOCR({ model: OCR_ENGLISH }); const runModel = async () => { const ocrDetections = await model.forward('https://url-to-image.jpg'); diff --git a/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md b/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md index 6f615a20c7..7449b0ea10 100644 --- a/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md +++ b/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md @@ -18,7 +18,7 @@ import { } from 'react-native-executorch'; function App() { - const ssdlite = useObjectDetection(SSDLITE_320_MOBILENET_V3_LARGE); + const ssdlite = useObjectDetection({ model: SSDLITE_320_MOBILENET_V3_LARGE }); // ... for (const detection of await ssdlite.forward('https://url-to-image.jpg')) { @@ -52,8 +52,10 @@ interface Detection { ### Arguments -**`modelSource`** - A string that specifies the path to the model file. You can download the model from our [HuggingFace repository](https://huggingface.co/software-mansion/react-native-executorch-ssdlite320-mobilenet-v3-large/tree/main). -For more information on that topic, you can check out the [Loading models](../../01-fundamentals/02-loading-models.md) page. +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the path to the model file. You can download the model from our [HuggingFace repository](https://huggingface.co/software-mansion/react-native-executorch-ssdlite320-mobilenet-v3-large/tree/main). + For more information on that topic, you can check out the [Loading models](../../01-fundamentals/02-loading-models.md) page. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -104,7 +106,7 @@ import { } from 'react-native-executorch'; function App() { - const ssdlite = useObjectDetection(SSDLITE_320_MOBILENET_V3_LARGE); + const ssdlite = useObjectDetection({ model: SSDLITE_320_MOBILENET_V3_LARGE }); const runModel = async () => { const detections = await ssdlite.forward('https://url-to-image.jpg'); diff --git a/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md b/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md index 44d4c31ba2..1688fc1972 100644 --- a/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md +++ b/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md @@ -16,7 +16,7 @@ import { STYLE_TRANSFER_CANDY, } from 'react-native-executorch'; -const model = useStyleTransfer(STYLE_TRANSFER_CANDY); +const model = useStyleTransfer({ model: STYLE_TRANSFER_CANDY }); const imageUri = 'file::///Users/.../cute_cat.png'; @@ -29,8 +29,9 @@ try { ### Arguments -**`modelSource`** -A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. @@ -56,7 +57,7 @@ Images from external sources and the generated image are stored in your applicat ```typescript function App() { - const model = useStyleTransfer(STYLE_TRANSFER_CANDY); + const model = useStyleTransfer({ model: STYLE_TRANSFER_CANDY }); // ... const imageUri = 'file::///Users/.../cute_cat.png'; diff --git a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md index ea285bc32a..148e31fe10 100644 --- a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md @@ -19,7 +19,7 @@ import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; function App() { const model = useVerticalOCR({ - ...VERTICAL_OCR_ENGLISH, + model: VERTICAL_OCR_ENGLISH, independentCharacters: true, }); @@ -129,19 +129,12 @@ interface OCRDetection { ### Arguments -**`detectorSources`** - An object that specifies the location of the detectors binary files. Each detector is composed of two models tailored to process images of varying widths. +**`model`** - Object containing the model source. -- `detectorLarge` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. -- `detectorNarrow` - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. - -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. - -**`recognizerSources`** - An object that specifies the locations of the recognizers binary files. Each recognizer is composed of two models tailored to process images of varying widths. - -- `recognizerLarge` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. -- `recognizerSmall` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 64 pixels. - -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +- **`detectorLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. +- **`detectorNarrow`** - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. +- **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. +- **`recognizerSmall`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 64 pixels. **`language`** - A parameter that specifies the language of the text to be recognized by the OCR. @@ -149,6 +142,8 @@ For more information, take a look at [loading models](../../01-fundamentals/02-l **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. + ### Returns The hook returns an object with the following properties: @@ -192,7 +187,7 @@ import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; function App() { const model = useVerticalOCR({ - ...VERTICAL_OCR_ENGLISH, + model: VERTICAL_OCR_ENGLISH, independentCharacters: true, }); diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md index a3aef792a8..92a9ea9da9 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md @@ -7,16 +7,7 @@ TypeScript API implementation of the [useLLM](../../02-hooks/01-natural-language ## Reference ```typescript -import { - LLAMA3_2_1B_QLORA, - LLAMA3_2_TOKENIZER, - LLAMA3_2_TOKENIZER_CONFIG, - LLMModule, -} from 'react-native-executorch'; - -const printDownloadProgress = (progress: number) => { - console.log(progress); -}; +import { LLMModule, LLAMA3_2_1B_QLORA } from 'react-native-executorch'; // Creating an instance const llm = new LLMModule({ @@ -25,12 +16,7 @@ const llm = new LLMModule({ }); // Loading the model -await llm.load({ - modelSource: LLAMA3_2_1B_QLORA, - tokenizerSource: LLAMA3_2_TOKENIZER, - tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG, - onDownloadProgressCallback: printDownloadProgress, -}); +await llm.load(LLAMA3_2_1B_QLORA, (progress) => console.log(progress)); // Running the model await llm.sendMessage('Hello, World!'); @@ -44,18 +30,18 @@ llm.delete(); ### Methods -| Method | Type | Description | -| ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `constructor` | `({tokenCallback?: (token: string) => void, responseCallback?: (response: string) => void, messageHistoryCallback?: (messageHistory: Message[]) => void})` | Creates a new instance of LLMModule with optional callbacks. | -| `load` | `({modelSource: ResourceSource, tokenizerSource: ResourceSource, tokenizerConfigSource: ResourceSource, onDownloadProgressCallback?: (downloadProgress: number) => void}) => Promise` | Loads the model. Checkout the [loading the model](#loading-the-model) section for details. | -| `setTokenCallback` | `{tokenCallback: (token: string) => void}) => void` | Sets new token callback. | -| `generate` | `(messages: Message[], tools?: LLMTool[]) => Promise` | Runs model to complete chat passed in `messages` argument. It doesn't manage conversation context. | -| `forward` | `(input: string) => Promise` | Runs model inference with raw input string. You need to provide entire conversation and prompt (in correct format and with special tokens!) in input string to this method. It doesn't manage conversation context. It is intended for users that need access to the model itself without any wrapper. If you want a simple chat with model the consider using`sendMessage` | -| `configure` | `({chatConfig?: Partial, toolsConfig?: ToolsConfig}) => void` | Configures chat and tool calling. See more details in [configuring the model](#configuring-the-model). | -| `sendMessage` | `(message: string) => Promise` | Method to add user message to conversation. After model responds it will call `messageHistoryCallback()`containing both user message and model response. It also returns them. | -| `deleteMessage` | `(index: number) => void` | Deletes all messages starting with message on `index` position. After deletion it will call `messageHistoryCallback()` containing new history. It also returns it. | -| `delete` | `() => void` | Method to delete the model from memory. Note you cannot delete model while it's generating. You need to interrupt it first and make sure model stopped generation. | -| `interrupt` | `() => void` | Interrupts model generation. It may return one more token after interrupt. | +| Method | Type | Description | +| ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `constructor` | `({tokenCallback?: (token: string) => void, responseCallback?: (response: string) => void, messageHistoryCallback?: (messageHistory: Message[]) => void})` | Creates a new instance of LLMModule with optional callbacks. | +| `load` | `(model: { modelSource: ResourceSource; tokenizerSource: ResourceSource; tokenizerConfigSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model. Checkout the [loading the model](#loading-the-model) section for details. | +| `setTokenCallback` | `{tokenCallback: (token: string) => void}) => void` | Sets new token callback. | +| `generate` | `(messages: Message[], tools?: LLMTool[]) => Promise` | Runs model to complete chat passed in `messages` argument. It doesn't manage conversation context. | +| `forward` | `(input: string) => Promise` | Runs model inference with raw input string. You need to provide entire conversation and prompt (in correct format and with special tokens!) in input string to this method. It doesn't manage conversation context. It is intended for users that need access to the model itself without any wrapper. If you want a simple chat with model the consider using`sendMessage` | +| `configure` | `({chatConfig?: Partial, toolsConfig?: ToolsConfig}) => void` | Configures chat and tool calling. See more details in [configuring the model](#configuring-the-model). | +| `sendMessage` | `(message: string) => Promise` | Method to add user message to conversation. After model responds it will call `messageHistoryCallback()`containing both user message and model response. It also returns them. | +| `deleteMessage` | `(index: number) => void` | Deletes all messages starting with message on `index` position. After deletion it will call `messageHistoryCallback()` containing new history. It also returns it. | +| `delete` | `() => void` | Method to delete the model from memory. Note you cannot delete model while it's generating. You need to interrupt it first and make sure model stopped generation. | +| `interrupt` | `() => void` | Interrupts model generation. It may return one more token after interrupt. |
Type definitions @@ -104,15 +90,17 @@ To create a new instance of LLMModule, use the constructor with optional callbac Then, to load the model, use the `load` method. It accepts an object with the following fields: -**`modelSource`** - A string that specifies the location of the model binary. +**`model`** - An object with the following fields: + +- **`modelSource`** - `ResourceSource` specifying the location of the model binary. -**`tokenizerSource`** - URL to the JSON file which contains the tokenizer. +- **`tokenizerSource`** - `ResourceSource` specifying the location of the tokenizer. -**`tokenizerConfigSource`** - URL to the JSON file which contains the tokenizer config. +- **`tokenizerConfigSource`** - `ResourceSource` specifying the location of the tokenizer config. **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -This method returns a promise, which can resolve to an error or void. It only works in managed chat (i.e. when you use `sendMessage`) +This method returns a promise, which can resolve to an error or void. ## Listening for download progress diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md index 7eed5211f9..6c0142c70c 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md @@ -33,7 +33,7 @@ const stt = new SpeechToTextModule({ // Loading the model await stt.load({ - modelName: 'moonshine', + model: MOONSHINE_TINY, onDownloadProgressCallback: (progress) => console.log(progress), }); @@ -44,15 +44,15 @@ const transcribedText = await stt.transcribe(waveform); ### Methods -| Method | Type | Description | -| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `constructor` | `({transcribeCallback?: (sequence: string) => void, overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality'})` | Creates a new instance of SpeechToTextModule with an optional transcription callback and streaming configuration. | -| `load` | `({modelName: 'whisper' \| 'moonshine' \| 'whisperMultilingual', encoderSource?: ResourceSource, decoderSource?: ResourceSource, tokenizerSource?: ResourceSource, onDownloadProgressCallback?: (downloadProgress: number) => void}) => Promise` | Loads the model specified with `modelName`, where `encoderSource`, `decoderSource`, `tokenizerSource` are strings specifying the location of the binaries for the models. `onDownloadProgressCallback` allows you to monitor the current progress of the model download | -| `transcribe` | `(waveform: number[], audioLanguage?: SpeechToTextLanguage): Promise` | Starts a transcription process for a given input array, which should be a waveform at 16kHz. Resolves a promise with the output transcription when the model is finished. For multilingual models, you have to specify the audioLanguage flag, which is the language of the spoken language in the audio. | -| `streamingTranscribe` | `(streamingAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | This allows for running transcription process on-line, which means where the whole audio is not known beforehand i.e. when transcribing from a live microphone feed. `streamingAction` defines the type of package sent to the model:
  • `START` - initializes the process, allows for optional `waveform` data
  • `DATA` - this package should contain consecutive audio data chunks sampled in 16k Hz
  • `STOP` - the last data chunk for this transcription, ends the transcription process and flushes internal buffers
    • Each call returns most recent transcription. Returns error when called when module is in use (i.e. processing `transcribe` call) | -| `encode` | `(waveform: number[]) => Promise` | Runs the encoding part of the model. It doesn't return the encodings. Instead, it stores the result internally, reducing data transfer overhead. | -| `decode` | `(tokens: number[]) => Promise` | Runs the decoder of the model. Returns a single token representing the next token in the output. It uses internal cached encodings from the most recent `encode` call, meaning that you have to call `encode` prior to decoding. | -| `configureStreaming` | `(overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality') => void` | Configures options for the streaming algorithm:
    • `overlapSeconds` determines how much adjacent audio chunks overlap (increasing it slows down transcription, decreases probability of weird wording at the chunks intersection, setting it larger than 3 seconds is generally discouraged),
    • `windowSize` describes size of the audio chunks (increasing it speeds up the end to end transcription time, but increases latency for the first token to be returned),
    • `streamingConfig` predefined configs for `windowSize` and `overlapSeconds` values.
    Keep `windowSize + 2 * overlapSeconds <= 30`. | +| Method | Type | Description | +| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `constructor` | `({transcribeCallback?: (sequence: string) => void, overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality'})` | Creates a new instance of SpeechToTextModule with an optional transcription callback and streaming configuration. | +| `load` | `(model: { modelName: AvailableModels; encoderSource?: ResourceSource; decoderSource?: ResourceSource; tokenizerSource?: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model specified with `modelName`, where `encoderSource`, `decoderSource`, `tokenizerSource` are strings specifying the location of the binaries for the models. `onDownloadProgressCallback` allows you to monitor the current progress of the model download | +| `transcribe` | `(waveform: number[], audioLanguage?: SpeechToTextLanguage): Promise` | Starts a transcription process for a given input array, which should be a waveform at 16kHz. Resolves a promise with the output transcription when the model is finished. For multilingual models, you have to specify the audioLanguage flag, which is the language of the spoken language in the audio. | +| `streamingTranscribe` | `(streamingAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | This allows for running transcription process on-line, which means where the whole audio is not known beforehand i.e. when transcribing from a live microphone feed. `streamingAction` defines the type of package sent to the model:
  • `START` - initializes the process, allows for optional `waveform` data
  • `DATA` - this package should contain consecutive audio data chunks sampled in 16k Hz
  • `STOP` - the last data chunk for this transcription, ends the transcription process and flushes internal buffers
    • Each call returns most recent transcription. Returns error when called when module is in use (i.e. processing `transcribe` call) | +| `encode` | `(waveform: number[]) => Promise` | Runs the encoding part of the model. It doesn't return the encodings. Instead, it stores the result internally, reducing data transfer overhead. | +| `decode` | `(tokens: number[]) => Promise` | Runs the decoder of the model. Returns a single token representing the next token in the output. It uses internal cached encodings from the most recent `encode` call, meaning that you have to call `encode` prior to decoding. | +| `configureStreaming` | `(overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality') => void` | Configures options for the streaming algorithm:
    • `overlapSeconds` determines how much adjacent audio chunks overlap (increasing it slows down transcription, decreases probability of weird wording at the chunks intersection, setting it larger than 3 seconds is generally discouraged),
    • `windowSize` describes size of the audio chunks (increasing it speeds up the end to end transcription time, but increases latency for the first token to be returned),
    • `streamingConfig` predefined configs for `windowSize` and `overlapSeconds` values.
    Keep `windowSize + 2 * overlapSeconds <= 30`. |
    Type definitions @@ -161,13 +161,15 @@ To create a new instance of SpeechToTextModule, use the constructor with optiona Then, to load the model, use the `load` method. It accepts an object with the following fields: -**`modelName`** - Identifier for which model to use ('whisper', 'moonshine', or 'whisperMultilingual'). +**`model`** - Object containing the model source. -**`encoderSource`** - (Optional) String that specifies the location of the encoder binary. +- **`modelName`** - Identifier for which model to use ('whisper', 'moonshine', or 'whisperMultilingual'). -**`decoderSource`** - (Optional) String that specifies the location of the decoder binary. +- **`encoderSource`** - (Optional) String that specifies the location of the encoder binary. -**`tokenizerSource`** - (Optional) String that specifies the location of the tokenizer. +- **`decoderSource`** - (Optional) String that specifies the location of the decoder binary. + +- **`tokenizerSource`** - (Optional) String that specifies the location of the tokenizer. **`onDownloadProgressCallback`** - (Optional) Function that will be called on download progress. diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md index 5c6b33c4b2..71da5b87ff 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md @@ -14,7 +14,9 @@ import { } from 'react-native-executorch'; // Loading the model -await TextEmbeddingsModule.load(ALL_MINILM_L6_V2, All_MINILM_L6_V2_TOKENIZER); +await TextEmbeddingsModule.load(ALL_MINILM_L6_V2, (progress) => + console.log(progress) +); // Running the model const embedding = await TextEmbeddingsModule.forward('Hello World!'); @@ -22,11 +24,11 @@ const embedding = await TextEmbeddingsModule.forward('Hello World!'); ### Methods -| Method | Type | Description | -| -------------------- | ------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `load` | `(modelSource: ResourceSource, tokenizerSource: ResourceSource): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary, `tokenizerSource` is a string that specifies the location of the tokenizer JSON file. | -| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` is a text that will be embedded. | -| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. | +| Method | Type | Description | +| -------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `load` | `(model: { modelSource: ResourceSource; tokenizerSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary, `tokenizerSource` is a string that specifies the location of the tokenizer JSON file. | +| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` is a text that will be embedded. | +| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. |
    Type definitions @@ -39,7 +41,14 @@ type ResourceSource = string | number | object; ## Loading the model -To load the model, use the `load` method. It accepts the `modelSource` which is a string that specifies the location of the model binary, `tokenizerSource` which is a string that specifies the location of the tokenizer JSON file. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, use the `load` method. It accepts an object: + +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. +- **`tokenizerSource`** - A string that specifies the location of the tokenizer JSON file. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md index a68f66b53f..728309356b 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md @@ -7,13 +7,10 @@ TypeScript API implementation of the [useTokenizer](../../02-hooks/01-natural-la ## Reference ```typescript -import { - TokenizerModule, - ALL_MINILM_L6_V2_TOKENIZER, -} from 'react-native-executorch'; +import { TokenizerModule, ALL_MINILM_L6_V2 } from 'react-native-executorch'; // Load the tokenizer -await TokenizerModule.load(ALL_MINILM_L6_V2_TOKENIZER); +await TokenizerModule.load(ALL_MINILM_L6_V2); console.log('Tokenizer loaded'); // Get tokenizers vocabulary size @@ -41,14 +38,14 @@ console.log('Token for ID:', token); ### Methods -| Method | Type | Description | -| -------------- | -------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------ | -| `load` | `(tokenizerSource: ResourceSource): Promise` | Loads the tokenizer from the specified source. `tokenizerSource` is a string that points to the location of the tokenizer JSON file. | -| `encode` | `(input: string): Promise` | Converts a string into an array of token IDs. | -| `decode` | `(input: number[]): Promise` | Converts an array of token IDs into a string. | -| `getVocabSize` | `(): Promise` | Returns the size of the tokenizer's vocabulary. | -| `idToToken` | `(tokenId: number): Promise` | Returns the token associated to the ID. | -| `tokenToId` | `(token: string): Promise` | Returns the ID associated to the token. | +| Method | Type | Description | +| -------------- | -------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------ | +| `load` | `(tokenizer: { tokenizerSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the tokenizer from the specified source. `tokenizerSource` is a string that points to the location of the tokenizer JSON file. | +| `encode` | `(input: string): Promise` | Converts a string into an array of token IDs. | +| `decode` | `(input: number[]): Promise` | Converts an array of token IDs into a string. | +| `getVocabSize` | `(): Promise` | Returns the size of the tokenizer's vocabulary. | +| `idToToken` | `(tokenId: number): Promise` | Returns the token associated to the ID. | +| `tokenToId` | `(token: string): Promise` | Returns the ID associated to the token. |
    Type definitions diff --git a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md index ca1cf99b58..5cdf575595 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md @@ -25,11 +25,11 @@ const classesWithProbabilities = await module.forward(imageUri); ### Methods -| Method | Type | Description | -| --------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -| `load` | `(modelSource: ResourceSource, onDownloadProgressCallback: (_: number) => void () => {}): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | -| `forward` | `(input: string): Promise<{ [category: string]: number }>` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | -| `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. | +| Method | Type | Description | +| --------- | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | +| `load` | `(model: { modelSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | +| `forward` | `(input: string): Promise<{ [category: string]: number }>` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | +| `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. |
    Type definitions @@ -42,7 +42,13 @@ type ResourceSource = string | number | object; ## Loading the model -To load the model, create a new instance of the module and use the `load` method on it. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, create a new instance of the module and use the `load` method on it. It accepts an object: + +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md index 04dd942db6..0c7d7b9e08 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md @@ -23,11 +23,11 @@ const embedding = await ImageEmbeddingsModule.forward( ### Methods -| Method | Type | Description | -| -------------------- | ----------------------------------------------------- | ------------------------------------------------------------------------------------------------- | -| `load` | `(modelSource: ResourceSource): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. | -| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` is a URI/URL to image that will be embedded. | -| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. | +| Method | Type | Description | +| -------------------- | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------- | +| `load` | `(model: { modelSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. | +| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` is a URI/URL to image that will be embedded. | +| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. |
    Type definitions @@ -40,7 +40,13 @@ type ResourceSource = string | number | object; ## Loading the model -To load the model, use the `load` method. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, use the `load` method. It accepts an object: + +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md index 7934717db6..92d786c0e1 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md @@ -27,7 +27,7 @@ const outputDict = await module.forward(imageUri); | Method | Type | Description | | --------- | ---------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `load` | `(modelSource: ResourceSource, onDownloadProgressCallback: (_: number) => void () => {}): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | +| `load` | `(model: { modelSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | | `forward` | `(input: string, classesOfInterest?: DeeplabLabel[], resize?: boolean) => Promise<{[key in DeeplabLabel]?: number[]}>` | Executes the model's forward pass, where :
    \* `input` can be a fetchable resource or a Base64-encoded string.
    \* `classesOfInterest` is an optional list of `DeeplabLabel` used to indicate additional arrays of probabilities to output (see section "Running the model"). The default is an empty list.
    \* `resize` is an optional boolean to indicate whether the output should be resized to the original image dimensions, or left in the size of the model (see section "Running the model"). The default is `false`.

    The return is a dictionary containing:
    \* for the key `DeeplabLabel.ARGMAX` an array of integers corresponding to the most probable class for each pixel
    \* an array of floats for each class from `classesOfInterest` corresponding to the probabilities for this class. | | `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. | @@ -42,7 +42,13 @@ type ResourceSource = string | number | object; ## Loading the model -To load the model, create a new instance of the module and use the `load` method on it. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, create a new instance of the module and use the `load` method on it. It accepts an object: + +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md index 4785d150a0..c1ebbd2252 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md @@ -7,25 +7,11 @@ TypeScript API implementation of the [useOCR](../../02-hooks/02-computer-vision/ ## Reference ```typescript -import { - OCRModule, - DETECTOR_CRAFT_800, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_256, - RECOGNIZER_EN_CRNN_128, -} from 'react-native-executorch'; +import { OCRModule, OCR_ENGLISH } from 'react-native-executorch'; const imageUri = 'path/to/image.png'; // Loading the model -await OCRModule.load({ - detectorSource: DETECTOR_CRAFT_800, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerMedium: RECOGNIZER_EN_CRNN_256, - recognizerSmall: RECOGNIZER_EN_CRNN_128, - }, - language: 'en', -}); +await OCRModule.load(OCR_ENGLISH); // Running the model const ocrDetections = await OCRModule.forward(imageUri); @@ -33,22 +19,16 @@ const ocrDetections = await OCRModule.forward(imageUri); ### Methods -| Method | Type | Description | -| -------------------- | ------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------- | -| `load` | `(detectorSource: string, recognizerSources: RecognizerSources, language: OCRLanguage): Promise` | Loads the detector and recognizers, which sources are represented by `RecognizerSources`. | -| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | -| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. | +| Method | Type | Description | +| -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `load` | `(model: { detectorSource: ResourceSource; recognizerLarge: ResourceSource; recognizerMedium: ResourceSource; recognizerSmall: ResourceSource; language: OCRLanguage }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `detectorSource` is a string that specifies the location of the detector binary, `recognizerLarge` is a string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels, `recognizerMedium` is a string that specifies the location of the recognizer binary file which accepts input images with a width of 256 pixels, `recognizerSmall` is a string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels, and `language` is a parameter that specifies the language of the text to be recognized by the OCR. | +| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | +| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. |
    Type definitions ```typescript -interface RecognizerSources { - recognizerLarge: string | number; - recognizerMedium: string | number; - recognizerSmall: string | number; -} - type OCRLanguage = | 'abq' | 'ady' @@ -131,21 +111,17 @@ interface OCRDetection { ## Loading the model -To load the model, use the `load` method. It accepts: - -**`detectorSource`** - A string that specifies the location of the detector binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. - -**`recognizerSources`** - An object that specifies locations of the recognizers binary files. Each recognizer is composed of three models tailored to process images of varying widths. - -- `recognizerLarge` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. -- `recognizerMedium` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 256 pixels. -- `recognizerSmall` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels. +To load the model, use the `load` method. It accepts an object: -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +**`model`** - Object containing the model source. -**`language`** - A parameter that specifies the language of the text to be recognized by the OCR. +- **`detectorSource`** - A string that specifies the location of the detector binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +- **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. +- **`recognizerMedium`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 256 pixels. +- **`recognizerSmall`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels. +- **`language`** - A parameter that specifies the language of the text to be recognized by the OCR. -This method returns a promise, which can resolve to an error or void. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. ## Listening for download progress diff --git a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md index 9b2ee878f2..cc2ba0c36c 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md @@ -25,11 +25,11 @@ const detections = await module.forward(imageUri); ### Methods -| Method | Type | Description | -| --------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -| `load` | `(modelSource: ResourceSource, onDownloadProgressCallback: (_: number) => void () => {}): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | -| `forward` | `(input: string, detectionThreshold: number = 0.7): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. `detectionThreshold` can be supplied to alter the sensitivity of the detection. | -| `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. | +| Method | Type | Description | +| --------- | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | +| `load` | `(model: { modelSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | +| `forward` | `(input: string, detectionThreshold: number = 0.7): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. `detectionThreshold` can be supplied to alter the sensitivity of the detection. | +| `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. |
    Type definitions @@ -55,7 +55,13 @@ interface Detection { ## Loading the model -To load the model, create a new instance of the module and use the `load` method on it. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, create a new instance of the module and use the `load` method on it. It accepts an object: + +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md index 0beabbb863..ac5273eb68 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md @@ -25,11 +25,11 @@ const generatedImageUrl = await module.forward(imageUri); ### Methods -| Method | Type | Description | -| --------- | -------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -| `load` | `(modelSource: ResourceSource, onDownloadProgressCallback: (_: number) => void () => {}): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | -| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | -| `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. | +| Method | Type | Description | +| --------- | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | +| `load` | `(model: { modelSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. To track the download progress, supply a callback function `onDownloadProgressCallback`. | +| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | +| `delete` | `(): void` | Release the memory held by the module. Calling `forward` afterwards is invalid. |
    Type definitions @@ -42,7 +42,13 @@ type ResourceSource = string | number | object; ## Loading the model -To load the model, create a new instance of the module and use the `load` method on it. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, create a new instance of the module and use the `load` method on it. It accepts an object: + +**`model`** - Object containing the model source. + +- **`modelSource`** - A string that specifies the location of the model binary. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md index da165b5120..ebdadb9197 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md @@ -7,29 +7,12 @@ TypeScript API implementation of the [useVerticalOCR](../../02-hooks/02-computer ## Reference ```typescript -import { - DETECTOR_CRAFT_1280, - DETECTOR_CRAFT_320, - RECOGNIZER_EN_CRNN_512, - RECOGNIZER_EN_CRNN_64, - useVerticalOCR, -} from 'react-native-executorch'; +import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; const imageUri = 'path/to/image.png'; // Loading the model -await VerticalOCRModule.load({ - detectorSources: { - detectorLarge: DETECTOR_CRAFT_1280, - detectorNarrow: DETECTOR_CRAFT_320, - }, - recognizerSources: { - recognizerLarge: RECOGNIZER_EN_CRNN_512, - recognizerSmall: RECOGNIZER_EN_CRNN_64, - }, - language: 'en', - independentCharacters: true, -}); +await VerticalOCRModule.load(VERTICAL_OCR_ENGLISH, true); // Running the model const ocrDetections = await VerticalOCRModule.forward(imageUri); @@ -37,11 +20,11 @@ const ocrDetections = await VerticalOCRModule.forward(imageUri); ### Methods -| Method | Type | Description | -| -------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ | -| `load` | `(detectorSources: DetectorSources, recognizerSources: RecognizerSources, language: OCRLanguage independentCharacters: boolean): Promise` | Loads detectors and recognizers, which sources are represented by `DetectorSources` and `RecognizerSources`. | -| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | -| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. | +| Method | Type | Description | +| -------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `load` | `(model: { detectorLarge: ResourceSource; detectorNarrow: ResourceSource; recognizerLarge: ResourceSource; recognizerSmall: ResourceSource; language: OCRLanguage }, independentCharacters: boolean, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `detectorLarge` is a string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels, `detectorNarrow` is a string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels, `recognizerLarge` is a string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels, `recognizerSmall` is a string that specifies the location of the recognizer binary file which accepts input images with a width of 64 pixels, and `language` is a parameter that specifies the language of the text to be recognized by the OCR. | +| `forward` | `(input: string): Promise` | Executes the model's forward pass, where `input` can be a fetchable resource or a Base64-encoded string. | +| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. |
    Type definitions @@ -141,25 +124,17 @@ interface OCRDetection { To load the model, use the `load` method. It accepts: -**`detectorSources`** - An object that specifies the location of the detectors binary files. Each detector is composed of two models tailored to process images of varying widths. +**`model`** - Object containing the model source. -- `detectorLarge` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. -- `detectorNarrow` - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. - -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. - -**`recognizerSources`** - An object that specifies the locations of the recognizers binary files. Each recognizer is composed of two models tailored to process images of varying widths. - -- `recognizerLarge` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. -- `recognizerSmall` - A string that specifies the location of the recognizer binary file which accepts input images with a width of 64 pixels. - -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. - -**`language`** - A parameter that specifies the language of the text to be recognized by the OCR. +- **`detectorLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. +- **`detectorNarrow`** - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. +- **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. +- **`recognizerSmall`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 64 pixels. +- **`language`** - A parameter that specifies the language of the text to be recognized by the OCR. **`independentCharacters`** – A boolean parameter that indicates whether the text in the image consists of a random sequence of characters. If set to true, the algorithm will scan each character individually instead of reading them as continuous text. -This method returns a promise, which can resolve to an error or void. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. ## Listening for download progress diff --git a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md index d7e9f2678a..fd2ef61319 100644 --- a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md +++ b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md @@ -25,13 +25,13 @@ const output = await ExecutorchModule.forward(input, shape); ### Methods -| Method | Type | Description | -| -------------------- | ------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `load` | `(modelSource: ResourceSource): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. | -| `forward` | `(input: ETInput, shape: number[]): Promise` | Executes the model's forward pass, where `input` is a JavaScript typed array and `shape` is an array of integers representing input Tensor shape. The output is a Tensor - raw result of inference. | -| `loadMethod` | `(methodName: string): Promise` | Loads resources specific to `methodName` into memory before execution. | -| `loadForward` | `(): Promise` | Loads resources specific to `forward` method into memory before execution. Uses `loadMethod` under the hood. | -| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. | +| Method | Type | Description | +| -------------------- | ------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `load` | `(modelSource: ResourceSource, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. | +| `forward` | `(input: ETInput, shape: number[]): Promise` | Executes the model's forward pass, where `input` is a JavaScript typed array and `shape` is an array of integers representing input Tensor shape. The output is a Tensor - raw result of inference. | +| `loadMethod` | `(methodName: string): Promise` | Loads resources specific to `methodName` into memory before execution. | +| `loadForward` | `(): Promise` | Loads resources specific to `forward` method into memory before execution. Uses `loadMethod` under the hood. | +| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. |
    Type definitions From 0c8de69790a9fe45b34df450c2c36b38773181fa Mon Sep 17 00:00:00 2001 From: jakmro Date: Thu, 31 Jul 2025 14:57:57 +0200 Subject: [PATCH 07/20] docs: add onDownloadProgressCallback parameter --- .../TextEmbeddingsModule.md | 2 ++ .../02-computer-vision/ClassificationModule.md | 2 ++ .../02-computer-vision/ImageEmbeddingsModule.md | 2 ++ .../02-computer-vision/ImageSegmentationModule.md | 2 ++ .../03-typescript-api/02-computer-vision/OCRModule.md | 6 ++---- .../02-computer-vision/ObjectDetectionModule.md | 2 ++ .../02-computer-vision/StyleTransferModule.md | 2 ++ .../02-computer-vision/VerticalOCRModule.md | 6 ++---- .../03-executorch-bindings/ExecutorchModule.md | 8 +++++++- 9 files changed, 23 insertions(+), 9 deletions(-) diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md index 71da5b87ff..36d8f8f48b 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md @@ -48,6 +48,8 @@ To load the model, use the `load` method. It accepts an object: - **`modelSource`** - A string that specifies the location of the model binary. - **`tokenizerSource`** - A string that specifies the location of the tokenizer JSON file. +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md index 5cdf575595..b2425e2a25 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md @@ -48,6 +48,8 @@ To load the model, create a new instance of the module and use the `load` method - **`modelSource`** - A string that specifies the location of the model binary. +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md index 0c7d7b9e08..ce8a16f1e6 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md @@ -46,6 +46,8 @@ To load the model, use the `load` method. It accepts an object: - **`modelSource`** - A string that specifies the location of the model binary. +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md index 92d786c0e1..7e4119d0b8 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md @@ -48,6 +48,8 @@ To load the model, create a new instance of the module and use the `load` method - **`modelSource`** - A string that specifies the location of the model binary. +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md index c1ebbd2252..37e86dce4c 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md @@ -121,11 +121,9 @@ To load the model, use the `load` method. It accepts an object: - **`recognizerSmall`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels. - **`language`** - A parameter that specifies the language of the text to be recognized by the OCR. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. - -## Listening for download progress +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. -To subscribe to the download progress event, you can use the `onDownloadProgress` method. It accepts a callback function that will be called whenever the download progress changes. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md index cc2ba0c36c..3dc8c816d0 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md @@ -61,6 +61,8 @@ To load the model, create a new instance of the module and use the `load` method - **`modelSource`** - A string that specifies the location of the model binary. +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md index ac5273eb68..e91b051564 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md @@ -48,6 +48,8 @@ To load the model, create a new instance of the module and use the `load` method - **`modelSource`** - A string that specifies the location of the model binary. +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md index ebdadb9197..4850e060a4 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md @@ -134,11 +134,9 @@ To load the model, use the `load` method. It accepts: **`independentCharacters`** – A boolean parameter that indicates whether the text in the image consists of a random sequence of characters. If set to true, the algorithm will scan each character individually instead of reading them as continuous text. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. - -## Listening for download progress +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. -To subscribe to the download progress event, you can use the `onDownloadProgress` method. It accepts a callback function that will be called whenever the download progress changes. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. ## Running the model diff --git a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md index fd2ef61319..92aa55bd52 100644 --- a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md +++ b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md @@ -51,7 +51,13 @@ export type ETInput = ## Loading the model -To load the model, use the `load` method. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +To load the model, use the `load` method. It accepts the following arguments: + +**`modelSource`** - A string that specifies the location of the model binary. + +**`onDownloadProgressCallback`** - (Optional) Function called on download progress. + +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model From f65100ef8d549ee38500d1bb8d68dd4f7a4af1d3 Mon Sep 17 00:00:00 2001 From: jakmro Date: Thu, 31 Jul 2025 15:22:31 +0200 Subject: [PATCH 08/20] refactor: update dependency array in useNonStaticModule hook --- .../react-native-executorch/src/hooks/useNonStaticModule.ts | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/packages/react-native-executorch/src/hooks/useNonStaticModule.ts b/packages/react-native-executorch/src/hooks/useNonStaticModule.ts index 05c15b347d..bebd41d187 100644 --- a/packages/react-native-executorch/src/hooks/useNonStaticModule.ts +++ b/packages/react-native-executorch/src/hooks/useNonStaticModule.ts @@ -51,7 +51,7 @@ export const useNonStaticModule = < }; // eslint-disable-next-line react-hooks/exhaustive-deps - }, [...Object.values(model), preventLoad]); + }, [_model, ...Object.values(model), preventLoad]); const forward = async (...input: ForwardArgs): Promise => { if (!isReady) throw new Error(getError(ETError.ModuleNotLoaded)); From 1b31761ab8a656e50e784467ad337c5e1a648f6d Mon Sep 17 00:00:00 2001 From: jakmro Date: Fri, 1 Aug 2025 07:33:09 +0200 Subject: [PATCH 09/20] docs: move references to loading models and remove export keyword --- docs/docs/02-hooks/01-natural-language-processing/useLLM.md | 6 ++++-- .../01-natural-language-processing/useSpeechToText.md | 4 +++- .../01-natural-language-processing/useTextEmbeddings.md | 4 +++- .../02-hooks/01-natural-language-processing/useTokenizer.md | 2 ++ .../03-executorch-bindings/ExecutorchModule.md | 2 +- 5 files changed, 13 insertions(+), 5 deletions(-) diff --git a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md index f5d16613a3..996b43bf8a 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md @@ -52,7 +52,7 @@ The code snippet above fetches the model from the specified URL, loads it into m **`model`** - Object containing the model source, tokenizer source, and tokenizer config source. -- **`modelSource`** - `ResourceSource` that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +- **`modelSource`** - `ResourceSource` that specifies the location of the model binary. - **`tokenizerSource`** - `ResourceSource` pointing to the JSON file which contains the tokenizer. @@ -60,6 +60,8 @@ The code snippet above fetches the model from the specified URL, loads it into m **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. + ### Returns | Field | Type | Description | @@ -93,7 +95,7 @@ const useLLM: ({ preventLoad?: boolean; }) => LLMType; -export interface LLMType { +interface LLMType { messageHistory: Message[]; response: string; token: string; diff --git a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md index 5bd06bef20..1ccbafeb32 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md @@ -70,7 +70,7 @@ Given that STT models can process audio no longer than 30 seconds, there is a ne - **`modelName`** - A literal of `"moonshine" | "whisper" | "whisperMultilingual` which serves as an identifier for which model should be used. -- **`encoderSource?`** - A string that specifies the location of a .pte file for the encoder. For further information on passing model sources, check out [Loading Models](../../01-fundamentals/02-loading-models.md). Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. +- **`encoderSource?`** - A string that specifies the location of a .pte file for the encoder. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. - **`decoderSource?`** - Analogous to the encoderSource, this takes in a string which is a source for the decoder part of the model. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. @@ -84,6 +84,8 @@ Given that STT models can process audio no longer than 30 seconds, there is a ne **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md index 768d4e4d09..a0e1340174 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md @@ -39,12 +39,14 @@ try { **`model`** - Object containing the model source and tokenizer source. -- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +- **`modelSource`** - A string that specifies the location of the model binary. - **`tokenizerSource`** - A string that specifies the location of the tokenizer JSON file. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md index eeb5bf3795..6126ee47f1 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md @@ -52,6 +52,8 @@ try { **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md index 92aa55bd52..5c08480e9f 100644 --- a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md +++ b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md @@ -39,7 +39,7 @@ const output = await ExecutorchModule.forward(input, shape); ```typescript type ResourceSource = string | number | object; -export type ETInput = +type ETInput = | Int8Array | Int32Array | BigInt64Array From 68d35a466a37f6a64b83942b35eb47076c437caf Mon Sep 17 00:00:00 2001 From: jakmro Date: Fri, 1 Aug 2025 08:00:17 +0200 Subject: [PATCH 10/20] feat: add AvailableModels export to sttDefaults and update index file --- .../react-native-executorch/src/constants/sttDefaults.ts | 2 ++ packages/react-native-executorch/src/index.tsx | 6 +++++- 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/packages/react-native-executorch/src/constants/sttDefaults.ts b/packages/react-native-executorch/src/constants/sttDefaults.ts index 655e37b6ff..586040a7ab 100644 --- a/packages/react-native-executorch/src/constants/sttDefaults.ts +++ b/packages/react-native-executorch/src/constants/sttDefaults.ts @@ -78,3 +78,5 @@ export enum STREAMING_ACTION { DATA, STOP, } + +export { AvailableModels }; diff --git a/packages/react-native-executorch/src/index.tsx b/packages/react-native-executorch/src/index.tsx index 2f38868e1d..48471df8e8 100644 --- a/packages/react-native-executorch/src/index.tsx +++ b/packages/react-native-executorch/src/index.tsx @@ -86,4 +86,8 @@ export { SpeechToTextLanguage }; export * from './constants/modelUrls'; export * from './constants/ocr/models'; export * from './constants/llmDefaults'; -export { STREAMING_ACTION, MODES } from './constants/sttDefaults'; +export { + STREAMING_ACTION, + MODES, + AvailableModels, +} from './constants/sttDefaults'; From bdfc9417b88c5c45fb4bb45def091d80b28fce7c Mon Sep 17 00:00:00 2001 From: jakmro Date: Fri, 1 Aug 2025 08:26:22 +0200 Subject: [PATCH 11/20] docs: update references to loading models to specify loading resources --- .../01-natural-language-processing/useLLM.md | 2 +- .../useSpeechToText.md | 2 +- .../useTextEmbeddings.md | 2 +- .../01-natural-language-processing/useTokenizer.md | 2 +- .../02-hooks/02-computer-vision/useClassification.md | 4 +++- .../02-computer-vision/useImageEmbeddings.md | 4 +++- .../02-computer-vision/useImageSegmentation.md | 4 +++- docs/docs/02-hooks/02-computer-vision/useOCR.md | 2 +- .../02-computer-vision/useObjectDetection.md | 3 ++- .../02-hooks/02-computer-vision/useStyleTransfer.md | 4 +++- .../02-hooks/02-computer-vision/useVerticalOCR.md | 2 +- .../03-executorch-bindings/useExecutorchModule.md | 4 +++- .../01-natural-language-processing/LLMModule.md | 4 +++- .../SpeechToTextModule.md | 12 ++++++++++-- .../TextEmbeddingsModule.md | 8 ++++---- .../02-computer-vision/ClassificationModule.md | 4 +++- .../02-computer-vision/ImageEmbeddingsModule.md | 4 +++- .../02-computer-vision/ImageSegmentationModule.md | 4 +++- .../02-computer-vision/OCRModule.md | 6 ++++-- .../02-computer-vision/ObjectDetectionModule.md | 4 +++- .../02-computer-vision/StyleTransferModule.md | 4 +++- .../02-computer-vision/VerticalOCRModule.md | 4 +++- .../03-executorch-bindings/ExecutorchModule.md | 4 +++- 23 files changed, 65 insertions(+), 28 deletions(-) diff --git a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md index 996b43bf8a..58ac30a4ee 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useLLM.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useLLM.md @@ -60,7 +60,7 @@ The code snippet above fetches the model from the specified URL, loads it into m **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Returns diff --git a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md index 1ccbafeb32..fd719f9d35 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md @@ -84,7 +84,7 @@ Given that STT models can process audio no longer than 30 seconds, there is a ne **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Returns diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md index a0e1340174..c40d19e94d 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTextEmbeddings.md @@ -45,7 +45,7 @@ try { **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Returns diff --git a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md index 6126ee47f1..23ad40803e 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useTokenizer.md @@ -52,7 +52,7 @@ try { **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Returns diff --git a/docs/docs/02-hooks/02-computer-vision/useClassification.md b/docs/docs/02-hooks/02-computer-vision/useClassification.md index 819aa0f7f0..cdb96f7327 100644 --- a/docs/docs/02-hooks/02-computer-vision/useClassification.md +++ b/docs/docs/02-hooks/02-computer-vision/useClassification.md @@ -32,10 +32,12 @@ try { **`model`** - Object containing the model source. -- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +- **`modelSource`** - A string that specifies the location of the model binary. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md b/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md index 8fa66d6802..96bf974c63 100644 --- a/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md +++ b/docs/docs/02-hooks/02-computer-vision/useImageEmbeddings.md @@ -43,10 +43,12 @@ try { **`model`** - Object containing the model source. -- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +- **`modelSource`** - A string that specifies the location of the model binary. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md b/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md index 031b0ea028..fa91ef37fc 100644 --- a/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md +++ b/docs/docs/02-hooks/02-computer-vision/useImageSegmentation.md @@ -31,10 +31,12 @@ try { **`model`** - Object containing the model source. -- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +- **`modelSource`** - A string that specifies the location of the model binary. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/02-hooks/02-computer-vision/useOCR.md b/docs/docs/02-hooks/02-computer-vision/useOCR.md index 62500c26ee..72974cf3f1 100644 --- a/docs/docs/02-hooks/02-computer-vision/useOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useOCR.md @@ -128,7 +128,7 @@ interface OCRDetection { **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Returns diff --git a/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md b/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md index 7449b0ea10..0eca74f5bb 100644 --- a/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md +++ b/docs/docs/02-hooks/02-computer-vision/useObjectDetection.md @@ -55,10 +55,11 @@ interface Detection { **`model`** - Object containing the model source. - **`modelSource`** - A string that specifies the path to the model file. You can download the model from our [HuggingFace repository](https://huggingface.co/software-mansion/react-native-executorch-ssdlite320-mobilenet-v3-large/tree/main). - For more information on that topic, you can check out the [Loading models](../../01-fundamentals/02-loading-models.md) page. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns The hook returns an object with the following properties: diff --git a/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md b/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md index 1688fc1972..a48b95cd7a 100644 --- a/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md +++ b/docs/docs/02-hooks/02-computer-vision/useStyleTransfer.md @@ -31,10 +31,12 @@ try { **`model`** - Object containing the model source. -- **`modelSource`** - A string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +- **`modelSource`** - A string that specifies the location of the model binary. **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ### Returns | Field | Type | Description | diff --git a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md index 148e31fe10..b273ef9490 100644 --- a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md @@ -142,7 +142,7 @@ interface OCRDetection { **`preventLoad?`** - Boolean that can prevent automatic model loading (and downloading the data if you load it for the first time) after running the hook. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Returns diff --git a/docs/docs/02-hooks/03-executorch-bindings/useExecutorchModule.md b/docs/docs/02-hooks/03-executorch-bindings/useExecutorchModule.md index e13673ed8b..55d5d8e2b8 100644 --- a/docs/docs/02-hooks/03-executorch-bindings/useExecutorchModule.md +++ b/docs/docs/02-hooks/03-executorch-bindings/useExecutorchModule.md @@ -20,7 +20,9 @@ const executorchModule = useExecutorchModule({ }); ``` -The `modelSource` parameter expects a location string pointing to the model binary. For more details on how to specify model sources, refer to the [loading models](../../01-fundamentals/02-loading-models.md) documentation. +The `modelSource` parameter expects a location string pointing to the model binary. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ### Arguments diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md index 92a9ea9da9..2d837d9593 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md @@ -33,7 +33,7 @@ llm.delete(); | Method | Type | Description | | ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `constructor` | `({tokenCallback?: (token: string) => void, responseCallback?: (response: string) => void, messageHistoryCallback?: (messageHistory: Message[]) => void})` | Creates a new instance of LLMModule with optional callbacks. | -| `load` | `(model: { modelSource: ResourceSource; tokenizerSource: ResourceSource; tokenizerConfigSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model. Checkout the [loading the model](#loading-the-model) section for details. | +| `load` | `(model: { modelSource: ResourceSource; tokenizerSource: ResourceSource; tokenizerConfigSource: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model. | | `setTokenCallback` | `{tokenCallback: (token: string) => void}) => void` | Sets new token callback. | | `generate` | `(messages: Message[], tools?: LLMTool[]) => Promise` | Runs model to complete chat passed in `messages` argument. It doesn't manage conversation context. | | `forward` | `(input: string) => Promise` | Runs model inference with raw input string. You need to provide entire conversation and prompt (in correct format and with special tokens!) in input string to this method. It doesn't manage conversation context. It is intended for users that need access to the model itself without any wrapper. If you want a simple chat with model the consider using`sendMessage` | @@ -102,6 +102,8 @@ Then, to load the model, use the `load` method. It accepts an object with the fo This method returns a promise, which can resolve to an error or void. +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. + ## Listening for download progress To subscribe to the download progress event, you can pass the `onDownloadProgressCallback` function to the `load` method. This function is called whenever the download progress changes. diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md index 6c0142c70c..2fda4efbec 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md @@ -66,6 +66,12 @@ enum STREAMING_ACTION { STOP, } +enum AvailableModels { + WHISPER = 'whisper', + MOONSHINE = 'moonshine', + WHISPER_MULTILINGUAL = 'whisperMultilingual', +} + enum SpeechToTextLanguage { Afrikaans = 'af', Albanian = 'sq', @@ -161,7 +167,7 @@ To create a new instance of SpeechToTextModule, use the constructor with optiona Then, to load the model, use the `load` method. It accepts an object with the following fields: -**`model`** - Object containing the model source. +**`model`** - Object containing the following fields: - **`modelName`** - Identifier for which model to use ('whisper', 'moonshine', or 'whisperMultilingual'). @@ -173,7 +179,9 @@ Then, to load the model, use the `load` method. It accepts an object with the fo **`onDownloadProgressCallback`** - (Optional) Function that will be called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md index 36d8f8f48b..fbcf652b28 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md @@ -14,9 +14,7 @@ import { } from 'react-native-executorch'; // Loading the model -await TextEmbeddingsModule.load(ALL_MINILM_L6_V2, (progress) => - console.log(progress) -); +await TextEmbeddingsModule.load(ALL_MINILM_L6_V2); // Running the model const embedding = await TextEmbeddingsModule.forward('Hello World!'); @@ -50,7 +48,9 @@ To load the model, use the `load` method. It accepts an object: **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md index b2425e2a25..e129a75732 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md @@ -50,7 +50,9 @@ To load the model, create a new instance of the module and use the `load` method **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md index ce8a16f1e6..4bb4033ebf 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md @@ -48,7 +48,9 @@ To load the model, use the `load` method. It accepts an object: **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md index 7e4119d0b8..ceaf2a524b 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md @@ -50,7 +50,9 @@ To load the model, create a new instance of the module and use the `load` method **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md index 37e86dce4c..9e5d8070a0 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md @@ -115,7 +115,7 @@ To load the model, use the `load` method. It accepts an object: **`model`** - Object containing the model source. -- **`detectorSource`** - A string that specifies the location of the detector binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. +- **`detectorSource`** - A string that specifies the location of the detector binary. - **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. - **`recognizerMedium`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 256 pixels. - **`recognizerSmall`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 128 pixels. @@ -123,7 +123,9 @@ To load the model, use the `load` method. It accepts an object: **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md index 3dc8c816d0..b68d6bbba9 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md @@ -63,7 +63,9 @@ To load the model, create a new instance of the module and use the `load` method **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md index e91b051564..3421dabf2b 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md @@ -50,7 +50,9 @@ To load the model, create a new instance of the module and use the `load` method **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md index 4850e060a4..f35c323246 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md @@ -136,7 +136,9 @@ To load the model, use the `load` method. It accepts: **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) section. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model diff --git a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md index 5c08480e9f..1c9fa7d540 100644 --- a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md +++ b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md @@ -57,7 +57,9 @@ To load the model, use the `load` method. It accepts the following arguments: **`onDownloadProgressCallback`** - (Optional) Function called on download progress. -For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. +This method returns a promise, which can resolve to an error or void. + +For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. ## Running the model From 868c300907409d69c8c4f5fe6c7843275dc94d29 Mon Sep 17 00:00:00 2001 From: jakmro Date: Fri, 1 Aug 2025 08:47:18 +0200 Subject: [PATCH 12/20] docs: clarify model object structure --- .../02-hooks/01-natural-language-processing/useSpeechToText.md | 2 +- docs/docs/02-hooks/02-computer-vision/useOCR.md | 2 +- docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md | 2 +- .../01-natural-language-processing/LLMModule.md | 2 +- .../01-natural-language-processing/SpeechToTextModule.md | 2 +- .../01-natural-language-processing/TextEmbeddingsModule.md | 2 +- docs/docs/03-typescript-api/02-computer-vision/OCRModule.md | 2 +- .../03-typescript-api/02-computer-vision/VerticalOCRModule.md | 2 +- 8 files changed, 8 insertions(+), 8 deletions(-) diff --git a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md index fd719f9d35..473d8fd9ce 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md @@ -68,7 +68,7 @@ Given that STT models can process audio no longer than 30 seconds, there is a ne **`model`** - Object containing the model name, encoder source, decoder source, and tokenizer source. -- **`modelName`** - A literal of `"moonshine" | "whisper" | "whisperMultilingual` which serves as an identifier for which model should be used. +- **`modelName`** - An enum `AvailableModels` value that serves as an identifier for which model should be used. - **`encoderSource?`** - A string that specifies the location of a .pte file for the encoder. Defaults to [constants](https://github.com/software-mansion/react-native-executorch/blob/main/packages/react-native-executorch/src/constants/modelUrls.ts) for given model. diff --git a/docs/docs/02-hooks/02-computer-vision/useOCR.md b/docs/docs/02-hooks/02-computer-vision/useOCR.md index 72974cf3f1..4afc63b2cf 100644 --- a/docs/docs/02-hooks/02-computer-vision/useOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useOCR.md @@ -118,7 +118,7 @@ interface OCRDetection { ### Arguments -**`model`** - Object containing the model source. +**`model`** - Object containing the detector source, recognizers sources, and language. - **`detectorSource`** - A string that specifies the location of the detector binary. - **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. diff --git a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md index b273ef9490..cc71282207 100644 --- a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md @@ -129,7 +129,7 @@ interface OCRDetection { ### Arguments -**`model`** - Object containing the model source. +**`model`** - Object containing the detectors sources, recognizers sources, and language. - **`detectorLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. - **`detectorNarrow`** - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md index 2d837d9593..14d3b4d748 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/LLMModule.md @@ -90,7 +90,7 @@ To create a new instance of LLMModule, use the constructor with optional callbac Then, to load the model, use the `load` method. It accepts an object with the following fields: -**`model`** - An object with the following fields: +**`model`** - Object containing the model source, tokenizer source, and tokenizer config source. - **`modelSource`** - `ResourceSource` specifying the location of the model binary. diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md index 2fda4efbec..1be6dd75a5 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md @@ -167,7 +167,7 @@ To create a new instance of SpeechToTextModule, use the constructor with optiona Then, to load the model, use the `load` method. It accepts an object with the following fields: -**`model`** - Object containing the following fields: +**`model`** - Object containing the model name, encoder source, decoder source, and tokenizer source. - **`modelName`** - Identifier for which model to use ('whisper', 'moonshine', or 'whisperMultilingual'). diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md index fbcf652b28..39ae24d29b 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md @@ -41,7 +41,7 @@ type ResourceSource = string | number | object; To load the model, use the `load` method. It accepts an object: -**`model`** - Object containing the model source. +**`model`** - Object containing the model source and tokenizer source. - **`modelSource`** - A string that specifies the location of the model binary. - **`tokenizerSource`** - A string that specifies the location of the tokenizer JSON file. diff --git a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md index 9e5d8070a0..436ca423e7 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md @@ -113,7 +113,7 @@ interface OCRDetection { To load the model, use the `load` method. It accepts an object: -**`model`** - Object containing the model source. +**`model`** - Object containing the detector source, recognizers sources, and language. - **`detectorSource`** - A string that specifies the location of the detector binary. - **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. diff --git a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md index f35c323246..f9a07d3858 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md @@ -124,7 +124,7 @@ interface OCRDetection { To load the model, use the `load` method. It accepts: -**`model`** - Object containing the model source. +**`model`** - Object containing the detectors sources, recognizers sources, and language. - **`detectorLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. - **`detectorNarrow`** - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. From 596c08e3c30b129fabaedd1b43b0115243a053ec Mon Sep 17 00:00:00 2001 From: jakmro Date: Fri, 1 Aug 2025 09:03:02 +0200 Subject: [PATCH 13/20] docs: correct model object descriptions --- docs/docs/02-hooks/02-computer-vision/useOCR.md | 2 +- docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md | 2 +- docs/docs/03-typescript-api/02-computer-vision/OCRModule.md | 2 +- .../03-typescript-api/02-computer-vision/VerticalOCRModule.md | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/docs/docs/02-hooks/02-computer-vision/useOCR.md b/docs/docs/02-hooks/02-computer-vision/useOCR.md index 4afc63b2cf..ad25998d76 100644 --- a/docs/docs/02-hooks/02-computer-vision/useOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useOCR.md @@ -118,7 +118,7 @@ interface OCRDetection { ### Arguments -**`model`** - Object containing the detector source, recognizers sources, and language. +**`model`** - Object containing the detector source, recognizer sources, and language. - **`detectorSource`** - A string that specifies the location of the detector binary. - **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. diff --git a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md index cc71282207..b449d9f07c 100644 --- a/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md +++ b/docs/docs/02-hooks/02-computer-vision/useVerticalOCR.md @@ -129,7 +129,7 @@ interface OCRDetection { ### Arguments -**`model`** - Object containing the detectors sources, recognizers sources, and language. +**`model`** - Object containing the detector sources, recognizer sources, and language. - **`detectorLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. - **`detectorNarrow`** - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. diff --git a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md index 436ca423e7..d2a389cd52 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md @@ -113,7 +113,7 @@ interface OCRDetection { To load the model, use the `load` method. It accepts an object: -**`model`** - Object containing the detector source, recognizers sources, and language. +**`model`** - Object containing the detector source, recognizer sources, and language. - **`detectorSource`** - A string that specifies the location of the detector binary. - **`recognizerLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 512 pixels. diff --git a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md index f9a07d3858..0d594c954b 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md @@ -124,7 +124,7 @@ interface OCRDetection { To load the model, use the `load` method. It accepts: -**`model`** - Object containing the detectors sources, recognizers sources, and language. +**`model`** - Object containing the detector sources, recognizer sources, and language. - **`detectorLarge`** - A string that specifies the location of the recognizer binary file which accepts input images with a width of 1280 pixels. - **`detectorNarrow`** - A string that specifies the location of the detector binary file which accepts input images with a width of 320 pixels. From 21f9178f6d70407e85dd1c37a38c6bb4fd40e04b Mon Sep 17 00:00:00 2001 From: jakmro Date: Mon, 4 Aug 2025 11:37:48 +0200 Subject: [PATCH 14/20] refactor: update module instantiation --- .../useSpeechToText.md | 20 +++++++++---------- .../SpeechToTextModule.md | 20 +++++++++---------- .../TextEmbeddingsModule.md | 8 +++++--- .../TokenizerModule.md | 15 ++++++++------ .../ClassificationModule.md | 7 ++++--- .../ImageEmbeddingsModule.md | 9 ++++++--- .../ImageSegmentationModule.md | 7 ++++--- .../02-computer-vision/OCRModule.md | 7 +++++-- .../ObjectDetectionModule.md | 7 ++++--- .../02-computer-vision/StyleTransferModule.md | 7 ++++--- .../02-computer-vision/VerticalOCRModule.md | 12 ++++++++--- .../ExecutorchModule.md | 7 +++++-- 12 files changed, 75 insertions(+), 51 deletions(-) diff --git a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md index 473d8fd9ce..d6a6914065 100644 --- a/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md +++ b/docs/docs/02-hooks/01-natural-language-processing/useSpeechToText.md @@ -88,16 +88,16 @@ For more information on loading resources, take a look at [loading models](../.. ### Returns -| Field | Type | Description | -| --------------------- | ------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `transcribe` | `(waveform: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | Starts a transcription process for a given input array, which should be a waveform at 16kHz. Resolves a promise with the output transcription when the model is finished. For multilingual models, you have to specify the audioLanguage flag, which is the language of the spoken language in the audio. Returns error when called when module is in use (i.e. in process of `streamingTranscribe` action) | -| `streamingTranscribe` | `(streamingAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | This allows for running transcription process on-line, which means where the whole audio is not known beforehand i.e. when transcribing from a live microphone feed. `streamingAction` defines the type of package sent to the model:
  • `START` - initializes the process, allows for optional `waveform` data
  • `DATA` - this package should contain consecutive audio data chunks sampled in 16k Hz
  • `STOP` - the last data chunk for this transcription, ends the transcription process and flushes internal buffers
    • Each call returns most recent transcription. Returns error when called when module is in use (i.e. processing `transcribe` call) | -| `error` | Error | undefined | Contains the error message if the model failed to load. | -| `sequence` | string | This property is updated with each generated token. If you're looking to obtain tokens as they're generated, you should use this property. | -| `isGenerating` | `boolean` | Indicates whether the model is currently processing an inference. | -| `isReady` | `boolean` | Indicates whether the model has successfully loaded and is ready for inference. | -| `configureStreaming` | (overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' | 'balanced' | 'quality') | Configures options for the streaming algorithm:
    • `overlapSeconds` determines how much adjacent audio chunks overlap (increasing it slows down transcription, decreases probability of weird wording at the chunks intersection, setting it larger than 3 seconds generally is discouraged),
    • `windowSize` describes size of the audio chunks (increasing it speeds up the end to end transcription time, but increases latency for the first token to be returned),
    • `streamingConfig` predefined configs for `windowSize` and `overlapSeconds` values.
    Keep `windowSize + 2 * overlapSeconds <= 30`. | -| `downloadProgress` | `number` | Tracks the progress of the model download process. | +| Field | Type | Description | +| --------------------- | ------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `transcribe` | `(waveform: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | Starts a transcription process for a given input array, which should be a waveform at 16kHz. Resolves a promise with the output transcription when the model is finished. For multilingual models, you have to specify the audioLanguage flag, which is the language of the spoken language in the audio. Returns error when called when module is in use (i.e. in process of `streamingTranscribe` action) | +| `streamingTranscribe` | `(streamingAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | This allows for running transcription process on-line, which means where the whole audio is not known beforehand i.e. when transcribing from a live microphone feed. `streamingAction` defines the type of package sent to the model:
    • `START` - initializes the process, allows for optional `waveform` data
    • `DATA` - this package should contain consecutive audio data chunks sampled in 16k Hz
    • `STOP` - the last data chunk for this transcription, ends the transcription process and flushes internal buffers
    Each call returns most recent transcription. Returns error when called when module is in use (i.e. processing `transcribe` call) | +| `error` | Error | undefined | Contains the error message if the model failed to load. | +| `sequence` | string | This property is updated with each generated token. If you're looking to obtain tokens as they're generated, you should use this property. | +| `isGenerating` | `boolean` | Indicates whether the model is currently processing an inference. | +| `isReady` | `boolean` | Indicates whether the model has successfully loaded and is ready for inference. | +| `configureStreaming` | (overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' | 'balanced' | 'quality') | Configures options for the streaming algorithm:
    • `overlapSeconds` determines how much adjacent audio chunks overlap (increasing it slows down transcription, decreases probability of weird wording at the chunks intersection, setting it larger than 3 seconds generally is discouraged),
    • `windowSize` describes size of the audio chunks (increasing it speeds up the end to end transcription time, but increases latency for the first token to be returned),
    • `streamingConfig` predefined configs for `windowSize` and `overlapSeconds` values.
    Keep `windowSize + 2 * overlapSeconds <= 30`. | +| `downloadProgress` | `number` | Tracks the progress of the model download process. |
    Type definitions diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md index 1be6dd75a5..bf465dc178 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md @@ -7,7 +7,7 @@ TypeScript API implementation of the [useSpeechToText](../../02-hooks/01-natural ## Reference ```typescript -import { SpeechToTextModule } from 'react-native-executorch'; +import { SpeechToTextModule, MOONSHINE_TINY } from 'react-native-executorch'; import { AudioContext } from 'react-native-audio-api'; import * as FileSystem from 'expo-file-system'; @@ -44,15 +44,15 @@ const transcribedText = await stt.transcribe(waveform); ### Methods -| Method | Type | Description | -| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `constructor` | `({transcribeCallback?: (sequence: string) => void, overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality'})` | Creates a new instance of SpeechToTextModule with an optional transcription callback and streaming configuration. | -| `load` | `(model: { modelName: AvailableModels; encoderSource?: ResourceSource; decoderSource?: ResourceSource; tokenizerSource?: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model specified with `modelName`, where `encoderSource`, `decoderSource`, `tokenizerSource` are strings specifying the location of the binaries for the models. `onDownloadProgressCallback` allows you to monitor the current progress of the model download | -| `transcribe` | `(waveform: number[], audioLanguage?: SpeechToTextLanguage): Promise` | Starts a transcription process for a given input array, which should be a waveform at 16kHz. Resolves a promise with the output transcription when the model is finished. For multilingual models, you have to specify the audioLanguage flag, which is the language of the spoken language in the audio. | -| `streamingTranscribe` | `(streamingAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | This allows for running transcription process on-line, which means where the whole audio is not known beforehand i.e. when transcribing from a live microphone feed. `streamingAction` defines the type of package sent to the model:
  • `START` - initializes the process, allows for optional `waveform` data
  • `DATA` - this package should contain consecutive audio data chunks sampled in 16k Hz
  • `STOP` - the last data chunk for this transcription, ends the transcription process and flushes internal buffers
    • Each call returns most recent transcription. Returns error when called when module is in use (i.e. processing `transcribe` call) | -| `encode` | `(waveform: number[]) => Promise` | Runs the encoding part of the model. It doesn't return the encodings. Instead, it stores the result internally, reducing data transfer overhead. | -| `decode` | `(tokens: number[]) => Promise` | Runs the decoder of the model. Returns a single token representing the next token in the output. It uses internal cached encodings from the most recent `encode` call, meaning that you have to call `encode` prior to decoding. | -| `configureStreaming` | `(overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality') => void` | Configures options for the streaming algorithm:
    • `overlapSeconds` determines how much adjacent audio chunks overlap (increasing it slows down transcription, decreases probability of weird wording at the chunks intersection, setting it larger than 3 seconds is generally discouraged),
    • `windowSize` describes size of the audio chunks (increasing it speeds up the end to end transcription time, but increases latency for the first token to be returned),
    • `streamingConfig` predefined configs for `windowSize` and `overlapSeconds` values.
    Keep `windowSize + 2 * overlapSeconds <= 30`. | +| Method | Type | Description | +| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `constructor` | `({transcribeCallback?: (sequence: string) => void, overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality'})` | Creates a new instance of SpeechToTextModule with an optional transcription callback and streaming configuration. | +| `load` | `(model: { modelName: AvailableModels; encoderSource?: ResourceSource; decoderSource?: ResourceSource; tokenizerSource?: ResourceSource }, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model specified with `modelName`, where `encoderSource`, `decoderSource`, `tokenizerSource` are strings specifying the location of the binaries for the models. `onDownloadProgressCallback` allows you to monitor the current progress of the model download | +| `transcribe` | `(waveform: number[], audioLanguage?: SpeechToTextLanguage): Promise` | Starts a transcription process for a given input array, which should be a waveform at 16kHz. Resolves a promise with the output transcription when the model is finished. For multilingual models, you have to specify the audioLanguage flag, which is the language of the spoken language in the audio. | +| `streamingTranscribe` | `(streamingAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage) => Promise` | This allows for running transcription process on-line, which means where the whole audio is not known beforehand i.e. when transcribing from a live microphone feed. `streamingAction` defines the type of package sent to the model:
    • `START` - initializes the process, allows for optional `waveform` data
    • `DATA` - this package should contain consecutive audio data chunks sampled in 16k Hz
    • `STOP` - the last data chunk for this transcription, ends the transcription process and flushes internal buffers
    Each call returns most recent transcription. Returns error when called when module is in use (i.e. processing `transcribe` call) | +| `encode` | `(waveform: number[]) => Promise` | Runs the encoding part of the model. It doesn't return the encodings. Instead, it stores the result internally, reducing data transfer overhead. | +| `decode` | `(tokens: number[]) => Promise` | Runs the decoder of the model. Returns a single token representing the next token in the output. It uses internal cached encodings from the most recent `encode` call, meaning that you have to call `encode` prior to decoding. | +| `configureStreaming` | `(overlapSeconds?: number, windowSize?: number, streamingConfig?: 'fast' \| 'balanced' \| 'quality') => void` | Configures options for the streaming algorithm:
    • `overlapSeconds` determines how much adjacent audio chunks overlap (increasing it slows down transcription, decreases probability of weird wording at the chunks intersection, setting it larger than 3 seconds is generally discouraged),
    • `windowSize` describes size of the audio chunks (increasing it speeds up the end to end transcription time, but increases latency for the first token to be returned),
    • `streamingConfig` predefined configs for `windowSize` and `overlapSeconds` values.
    Keep `windowSize + 2 * overlapSeconds <= 30`. |
    Type definitions diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md index 39ae24d29b..7f59268f97 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TextEmbeddingsModule.md @@ -10,14 +10,16 @@ TypeScript API implementation of the [useTextEmbeddings](../../02-hooks/01-natur import { TextEmbeddingsModule, ALL_MINILM_L6_V2, - All_MINILM_L6_V2_TOKENIZER, } from 'react-native-executorch'; +// Creating an instance +const textEmbeddingsModule = new TextEmbeddingsModule(); + // Loading the model -await TextEmbeddingsModule.load(ALL_MINILM_L6_V2); +await textEmbeddingsModule.load(ALL_MINILM_L6_V2); // Running the model -const embedding = await TextEmbeddingsModule.forward('Hello World!'); +const embedding = await textEmbeddingsModule.forward('Hello World!'); ``` ### Methods diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md index 728309356b..41ad2b0279 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/TokenizerModule.md @@ -9,30 +9,33 @@ TypeScript API implementation of the [useTokenizer](../../02-hooks/01-natural-la ```typescript import { TokenizerModule, ALL_MINILM_L6_V2 } from 'react-native-executorch'; +// Creating an instance +const tokenizerModule = new TokenizerModule(); + // Load the tokenizer -await TokenizerModule.load(ALL_MINILM_L6_V2); +await tokenizerModule.load(ALL_MINILM_L6_V2); console.log('Tokenizer loaded'); // Get tokenizers vocabulary size -const vocabSize = await TokenizerModule.getVocabSize(); +const vocabSize = await tokenizerModule.getVocabSize(); console.log('Vocabulary size:', vocabSize); const text = 'Hello, world!'; // Tokenize the text -const tokens = await TokenizerModule.encode(text); +const tokens = await tokenizerModule.encode(text); console.log('Token IDs:', tokens); // Decode the tokens back to text -const decoded = await TokenizerModule.decode(tokens); +const decoded = await tokenizerModule.decode(tokens); console.log('Decoded text:', decoded); // Get the token ID for a specific token -const tokenId = await TokenizerModule.tokenToId('hello'); +const tokenId = await tokenizerModule.tokenToId('hello'); console.log('Token ID for "Hello":', tokenId); // Get the token for a specific ID -const token = await TokenizerModule.idToToken(tokenId); +const token = await tokenizerModule.idToToken(tokenId); console.log('Token for ID:', token); ``` diff --git a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md index e129a75732..7ddf0f0183 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ClassificationModule.md @@ -14,13 +14,14 @@ import { const imageUri = 'path/to/image.png'; -const module = new ClassificationModule(); +// Creating an instance +const classificationModule = new ClassificationModule(); // Loading the model -await module.load(EFFICIENTNET_V2_S); +await classificationModule.load(EFFICIENTNET_V2_S); // Running the model -const classesWithProbabilities = await module.forward(imageUri); +const classesWithProbabilities = await classificationModule.forward(imageUri); ``` ### Methods diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md index 4bb4033ebf..2333f57b49 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageEmbeddingsModule.md @@ -9,14 +9,17 @@ TypeScript API implementation of the [useImageEmbeddings](../../02-hooks/02-comp ```typescript import { ImageEmbeddingsModule, - CLIP_VIT_BASE_PATCH_32_IMAGE_ENCODER, + CLIP_VIT_BASE_PATCH32_IMAGE, } from 'react-native-executorch'; +// Creating an instance +const imageEmbeddingsModule = new ImageEmbeddingsModule(); + // Loading the model -await ImageEmbeddingsModule.load(CLIP_VIT_BASE_PATCH_32_IMAGE_ENCODER); +await imageEmbeddingsModule.load(CLIP_VIT_BASE_PATCH32_IMAGE); // Running the model -const embedding = await ImageEmbeddingsModule.forward( +const embedding = await imageEmbeddingsModule.forward( 'https://url-to-image.jpg' ); ``` diff --git a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md index ceaf2a524b..9d4e1e9291 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ImageSegmentationModule.md @@ -14,13 +14,14 @@ import { const imageUri = 'path/to/image.png'; -const module = new ImageSegmentationModule(); +// Creating an instance +const imageSegmentationModule = new ImageSegmentationModule(); // Loading the model -await module.load(DEEPLAB_V3_RESNET50); +await imageSegmentationModule.load(DEEPLAB_V3_RESNET50); // Running the model -const outputDict = await module.forward(imageUri); +const outputDict = await imageSegmentationModule.forward(imageUri); ``` ### Methods diff --git a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md index d2a389cd52..f709ffe1a4 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/OCRModule.md @@ -10,11 +10,14 @@ TypeScript API implementation of the [useOCR](../../02-hooks/02-computer-vision/ import { OCRModule, OCR_ENGLISH } from 'react-native-executorch'; const imageUri = 'path/to/image.png'; +// Creating an instance +const ocrModule = new OCRModule(); + // Loading the model -await OCRModule.load(OCR_ENGLISH); +await ocrModule.load(OCR_ENGLISH); // Running the model -const ocrDetections = await OCRModule.forward(imageUri); +const detections = await ocrModule.forward(imageUri); ``` ### Methods diff --git a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md index b68d6bbba9..07506a5c57 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/ObjectDetectionModule.md @@ -14,13 +14,14 @@ import { const imageUri = 'path/to/image.png'; -const module = new ObjectDetectionModule(); +// Creating an instance +const objectDetectionModule = new ObjectDetectionModule(); // Loading the model -await module.load(SSDLITE_320_MOBILENET_V3_LARGE); +await objectDetectionModule.load(SSDLITE_320_MOBILENET_V3_LARGE); // Running the model -const detections = await module.forward(imageUri); +const detections = await objectDetectionModule.forward(imageUri); ``` ### Methods diff --git a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md index 3421dabf2b..87ba6c955d 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/StyleTransferModule.md @@ -14,13 +14,14 @@ import { const imageUri = 'path/to/image.png'; -const module = new StyleTransferModule(); +// Creating an instance +const styleTransferModule = new StyleTransferModule(); // Loading the model -await module.load(STYLE_TRANSFER_CANDY); +await styleTransferModule.load(STYLE_TRANSFER_CANDY); // Running the model -const generatedImageUrl = await module.forward(imageUri); +const generatedImageUrl = await styleTransferModule.forward(imageUri); ``` ### Methods diff --git a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md index 0d594c954b..bf3b56c7e5 100644 --- a/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md +++ b/docs/docs/03-typescript-api/02-computer-vision/VerticalOCRModule.md @@ -7,15 +7,21 @@ TypeScript API implementation of the [useVerticalOCR](../../02-hooks/02-computer ## Reference ```typescript -import { useVerticalOCR, VERTICAL_OCR_ENGLISH } from 'react-native-executorch'; +import { + VerticalOCRModule, + VERTICAL_OCR_ENGLISH, +} from 'react-native-executorch'; const imageUri = 'path/to/image.png'; +// Creating an instance +const verticalOCRModule = new VerticalOCRModule(); + // Loading the model -await VerticalOCRModule.load(VERTICAL_OCR_ENGLISH, true); +await verticalOCRModule.load(VERTICAL_OCR_ENGLISH); // Running the model -const ocrDetections = await VerticalOCRModule.forward(imageUri); +const detections = await verticalOCRModule.forward(imageUri); ``` ### Methods diff --git a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md index 1c9fa7d540..958753291b 100644 --- a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md +++ b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md @@ -16,11 +16,14 @@ import { const shape = [1, 3, 640, 640]; const input = new Float32Array(1 * 3 * 640 * 640); +// Creating an instance +const executorchModule = new ExecutorchModule(); + // Loading the model -await ExecutorchModule.load(STYLE_TRANSFER_CANDY); +await executorchModule.load(STYLE_TRANSFER_CANDY); // Running the model -const output = await ExecutorchModule.forward(input, shape); +const output = await executorchModule.forward(input, shape); ``` ### Methods From 62c10190bf6a1e80d50d688e05e19bc8de840415 Mon Sep 17 00:00:00 2001 From: jakmro Date: Mon, 4 Aug 2025 11:56:04 +0200 Subject: [PATCH 15/20] style: format code --- .../src/controllers/SpeechToTextController.ts | 14 +++++++------- .../natural_language_processing/useTokenizer.ts | 5 ++++- .../TextEmbeddingsModule.ts | 7 +++++-- 3 files changed, 16 insertions(+), 10 deletions(-) diff --git a/packages/react-native-executorch/src/controllers/SpeechToTextController.ts b/packages/react-native-executorch/src/controllers/SpeechToTextController.ts index e773c17547..3661b9938f 100644 --- a/packages/react-native-executorch/src/controllers/SpeechToTextController.ts +++ b/packages/react-native-executorch/src/controllers/SpeechToTextController.ts @@ -100,9 +100,9 @@ export class SpeechToTextController { this.config = MODEL_CONFIGS[modelName]; try { - const tokenizerLoadPromise = this.tokenizerModule.load( - { tokenizerSource: tokenizerSource || this.config.tokenizer.source } - ); + const tokenizerLoadPromise = this.tokenizerModule.load({ + tokenizerSource: tokenizerSource || this.config.tokenizer.source, + }); const pathsPromise = ResourceFetcher.fetch( onDownloadProgressCallback, encoderSource || this.config.sources.encoder, @@ -414,7 +414,7 @@ export class SpeechToTextController { const chunk = this.waveform.slice( 0, this.windowSize + - this.overlapSeconds * (1 + Number(this.seqs.length > 0)) + this.overlapSeconds * (1 + Number(this.seqs.length > 0)) ); this.chunks = [chunk]; //save last chunk for STREAMING_ACTION.STOP this.waveform = this.waveform.slice( @@ -435,9 +435,9 @@ export class SpeechToTextController { // pad remaining waveform data with previous chunk to this.windowSize + 2 * this.overlapSeconds const chunk = this.chunks.length ? [ - ...this.chunks[0]!.slice(0, this.windowSize), - ...this.waveform, - ].slice(-this.windowSize - 2 * this.overlapSeconds) + ...this.chunks[0]!.slice(0, this.windowSize), + ...this.waveform, + ].slice(-this.windowSize - 2 * this.overlapSeconds) : this.waveform; this.waveform = []; diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts index 5101bbb8b4..3c01f81fdb 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useTokenizer.ts @@ -23,7 +23,10 @@ export const useTokenizer = ({ setError(null); try { setIsReady(false); - await _tokenizer.load({ tokenizerSource: tokenizer.tokenizerSource }, setDownloadProgress); + await _tokenizer.load( + { tokenizerSource: tokenizer.tokenizerSource }, + setDownloadProgress + ); setIsReady(true); } catch (err) { setError((err as Error).message); diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts index ac73e27404..db19ecbb66 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/TextEmbeddingsModule.ts @@ -5,13 +5,16 @@ import { BaseNonStaticModule } from '../BaseNonStaticModule'; export class TextEmbeddingsModule extends BaseNonStaticModule { async load( model: { modelSource: ResourceSource; tokenizerSource: ResourceSource }, - onDownloadProgressCallback: (progress: number) => void = () => { } + onDownloadProgressCallback: (progress: number) => void = () => {} ): Promise { const modelPromise = ResourceFetcher.fetch( onDownloadProgressCallback, model.modelSource ); - const tokenizerPromise = ResourceFetcher.fetch(undefined, model.tokenizerSource); + const tokenizerPromise = ResourceFetcher.fetch( + undefined, + model.tokenizerSource + ); const [modelResult, tokenizerResult] = await Promise.all([ modelPromise, tokenizerPromise, From ad0f59b3cf9fe14f9c6491f9a32551f2019110f9 Mon Sep 17 00:00:00 2001 From: jakmro Date: Mon, 4 Aug 2025 12:14:03 +0200 Subject: [PATCH 16/20] docs: change model loading syntax in SpeechToTextModule documentation --- .../01-natural-language-processing/SpeechToTextModule.md | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md index bf465dc178..104c78b0b0 100644 --- a/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md +++ b/docs/docs/03-typescript-api/01-natural-language-processing/SpeechToTextModule.md @@ -32,10 +32,7 @@ const stt = new SpeechToTextModule({ }); // Loading the model -await stt.load({ - model: MOONSHINE_TINY, - onDownloadProgressCallback: (progress) => console.log(progress), -}); +await stt.load(MOONSHINE_TINY, (progress) => console.log(progress)); // Loading the audio and running the model const waveform = await loadAudio(audioUrl); From 24058aeaba432f2dddca309b4f53ae71e8317613 Mon Sep 17 00:00:00 2001 From: jakmro Date: Tue, 5 Aug 2025 10:27:14 +0200 Subject: [PATCH 17/20] refactor: add requested changes --- .../src/constants/ocr/models.ts | 14 ++++++++------ .../hooks/natural_language_processing/useLLM.ts | 4 +--- .../natural_language_processing/LLMModule.ts | 4 +--- 3 files changed, 10 insertions(+), 12 deletions(-) diff --git a/packages/react-native-executorch/src/constants/ocr/models.ts b/packages/react-native-executorch/src/constants/ocr/models.ts index 16b7c7654c..18ce2500a3 100644 --- a/packages/react-native-executorch/src/constants/ocr/models.ts +++ b/packages/react-native-executorch/src/constants/ocr/models.ts @@ -1,4 +1,4 @@ -import { symbols } from './symbols'; +import { alphabets, symbols } from './symbols'; const URL_PREFIX = 'https://huggingface.co/software-mansion/react-native-executorch'; @@ -8,7 +8,9 @@ const DETECTOR_CRAFT_1280_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/x const DETECTOR_CRAFT_800_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_800.pte`; const DETECTOR_CRAFT_320_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/xnnpack/xnnpack_craft_320.pte`; -const createHFRecognizerDownloadUrl = (alphabet: string, size: number) => +type RecognizerSize = 64 | 128 | 256 | 512; + +const createHFRecognizerDownloadUrl = (alphabet: keyof typeof alphabets, size: RecognizerSize) => `${URL_PREFIX}-recognizer-crnn.en/${VERSION_TAG}/xnnpack/${alphabet}/xnnpack_crnn_${alphabet}_${size}.pte`; const RECOGNIZER_ENGLISH_CRNN_512 = createHFRecognizerDownloadUrl( @@ -71,10 +73,10 @@ const RECOGNIZER_TELUGU_CRNN_256 = createHFRecognizerDownloadUrl('telugu', 256); const RECOGNIZER_TELUGU_CRNN_128 = createHFRecognizerDownloadUrl('telugu', 128); const RECOGNIZER_TELUGU_CRNN_64 = createHFRecognizerDownloadUrl('telugu', 64); -const RECOGNIZER_ZH_SIM_CRNN_512 = createHFRecognizerDownloadUrl('zh-sim', 512); -const RECOGNIZER_ZH_SIM_CRNN_256 = createHFRecognizerDownloadUrl('zh-sim', 256); -const RECOGNIZER_ZH_SIM_CRNN_128 = createHFRecognizerDownloadUrl('zh-sim', 128); -const RECOGNIZER_ZH_SIM_CRNN_64 = createHFRecognizerDownloadUrl('zh-sim', 64); +const RECOGNIZER_ZH_SIM_CRNN_512 = createHFRecognizerDownloadUrl('zhSim', 512); +const RECOGNIZER_ZH_SIM_CRNN_256 = createHFRecognizerDownloadUrl('zhSim', 256); +const RECOGNIZER_ZH_SIM_CRNN_128 = createHFRecognizerDownloadUrl('zhSim', 128); +const RECOGNIZER_ZH_SIM_CRNN_64 = createHFRecognizerDownloadUrl('zhSim', 64); const RECOGNIZER_CYRILLIC_CRNN_512 = createHFRecognizerDownloadUrl( 'cyrillic', diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts index ee6d32f955..6872f6595e 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts @@ -56,9 +56,7 @@ export const useLLM = ({ (async () => { try { await _model.load({ - modelSource: model.modelSource, - tokenizerSource: model.tokenizerSource, - tokenizerConfigSource: model.tokenizerConfigSource, + ...model, onDownloadProgressCallback: setDownloadProgress, }); } catch (e) { diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts index 8c5ebdbdb7..2993308078 100644 --- a/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts +++ b/packages/react-native-executorch/src/modules/natural_language_processing/LLMModule.ts @@ -30,9 +30,7 @@ export class LLMModule { onDownloadProgressCallback: (progress: number) => void = () => {} ) { await this.controller.load({ - modelSource: model.modelSource, - tokenizerSource: model.tokenizerSource, - tokenizerConfigSource: model.tokenizerConfigSource, + ...model, onDownloadProgressCallback, }); } From 9b90a53732fae604af038c9562520eb03aaf56a3 Mon Sep 17 00:00:00 2001 From: jakmro Date: Tue, 5 Aug 2025 10:53:58 +0200 Subject: [PATCH 18/20] refactor: rename model instances --- .../src/hooks/computer_vision/useOCR.ts | 8 ++--- .../hooks/computer_vision/useVerticalOCR.ts | 8 ++--- .../natural_language_processing/useLLM.ts | 33 +++++++++++-------- .../useSpeechToText.ts | 25 +++++++++----- .../src/hooks/useNonStaticModule.ts | 10 +++--- 5 files changed, 49 insertions(+), 35 deletions(-) diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts b/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts index 0b0020b054..1aa5a2e51c 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useOCR.ts @@ -29,7 +29,7 @@ export const useOCR = ({ const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const _model = useMemo( + const controllerInstance = useMemo( () => new OCRController({ modelDownloadProgressCallback: setDownloadProgress, @@ -42,7 +42,7 @@ export const useOCR = ({ useEffect(() => { const loadModel = async () => { - await _model.loadModel( + await controllerInstance.loadModel( model.detectorSource, { recognizerLarge: model.recognizerLarge, @@ -57,7 +57,7 @@ export const useOCR = ({ loadModel(); } }, [ - _model, + controllerInstance, model.detectorSource, model.recognizerLarge, model.recognizerMedium, @@ -70,7 +70,7 @@ export const useOCR = ({ error, isReady, isGenerating, - forward: _model.forward, + forward: controllerInstance.forward, downloadProgress, }; }; diff --git a/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts b/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts index cccdc58477..31ea5832d8 100644 --- a/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts +++ b/packages/react-native-executorch/src/hooks/computer_vision/useVerticalOCR.ts @@ -31,7 +31,7 @@ export const useVerticalOCR = ({ const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const _model = useMemo( + const controllerInstance = useMemo( () => new VerticalOCRController({ modelDownloadProgressCallback: setDownloadProgress, @@ -46,7 +46,7 @@ export const useVerticalOCR = ({ if (preventLoad) return; (async () => { - await _model.loadModel( + await controllerInstance.loadModel( { detectorLarge: model.detectorLarge, detectorNarrow: model.detectorNarrow, @@ -60,7 +60,7 @@ export const useVerticalOCR = ({ ); })(); }, [ - _model, + controllerInstance, model.detectorLarge, model.detectorNarrow, model.recognizerLarge, @@ -74,7 +74,7 @@ export const useVerticalOCR = ({ error, isReady, isGenerating, - forward: _model.forward, + forward: controllerInstance.forward, downloadProgress, }; }; diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts index 6872f6595e..967d9913ee 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useLLM.ts @@ -36,7 +36,7 @@ export const useLLM = ({ setResponse((prevResponse) => prevResponse + newToken); }, []); - const _model = useMemo( + const controllerInstance = useMemo( () => new LLMController({ tokenCallback: tokenCallback, @@ -55,8 +55,10 @@ export const useLLM = ({ (async () => { try { - await _model.load({ - ...model, + await controllerInstance.load({ + modelSource: model.modelSource, + tokenizerSource: model.tokenizerSource, + tokenizerConfigSource: model.tokenizerConfigSource, onDownloadProgressCallback: setDownloadProgress, }); } catch (e) { @@ -65,10 +67,10 @@ export const useLLM = ({ })(); return () => { - _model.delete(); + controllerInstance.delete(); }; }, [ - _model, + controllerInstance, model.modelSource, model.tokenizerSource, model.tokenizerConfigSource, @@ -83,32 +85,35 @@ export const useLLM = ({ }: { chatConfig?: Partial; toolsConfig?: ToolsConfig; - }) => _model.configure({ chatConfig, toolsConfig }), - [_model] + }) => controllerInstance.configure({ chatConfig, toolsConfig }), + [controllerInstance] ); const generate = useCallback( (messages: Message[], tools?: LLMTool[]) => { setResponse(''); - return _model.generate(messages, tools); + return controllerInstance.generate(messages, tools); }, - [_model] + [controllerInstance] ); const sendMessage = useCallback( (message: string) => { setResponse(''); - return _model.sendMessage(message); + return controllerInstance.sendMessage(message); }, - [_model] + [controllerInstance] ); const deleteMessage = useCallback( - (index: number) => _model.deleteMessage(index), - [_model] + (index: number) => controllerInstance.deleteMessage(index), + [controllerInstance] ); - const interrupt = useCallback(() => _model.interrupt(), [_model]); + const interrupt = useCallback( + () => controllerInstance.interrupt(), + [controllerInstance] + ); return { messageHistory, diff --git a/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts b/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts index 1bb493e105..d1549bf70f 100644 --- a/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts +++ b/packages/react-native-executorch/src/hooks/natural_language_processing/useSpeechToText.ts @@ -52,7 +52,7 @@ export const useSpeechToText = ({ const [isGenerating, setIsGenerating] = useState(false); const [error, setError] = useState(); - const _model = useMemo( + const controllerInstance = useMemo( () => new SpeechToTextController({ transcribeCallback: setSequence, @@ -64,12 +64,16 @@ export const useSpeechToText = ({ ); useEffect(() => { - _model.configureStreaming(overlapSeconds, windowSize, streamingConfig); - }, [_model, overlapSeconds, windowSize, streamingConfig]); + controllerInstance.configureStreaming( + overlapSeconds, + windowSize, + streamingConfig + ); + }, [controllerInstance, overlapSeconds, windowSize, streamingConfig]); useEffect(() => { const loadModel = async () => { - await _model.load({ + await controllerInstance.load({ modelName: model.modelName, encoderSource: model.encoderSource, decoderSource: model.decoderSource, @@ -81,7 +85,7 @@ export const useSpeechToText = ({ loadModel(); } }, [ - _model, + controllerInstance, model.modelName, model.encoderSource, model.decoderSource, @@ -93,15 +97,20 @@ export const useSpeechToText = ({ isReady, isGenerating, downloadProgress, - configureStreaming: _model.configureStreaming, + configureStreaming: controllerInstance.configureStreaming, sequence, error, transcribe: (waveform: number[], audioLanguage?: SpeechToTextLanguage) => - _model.transcribe(waveform, audioLanguage), + controllerInstance.transcribe(waveform, audioLanguage), streamingTranscribe: ( streamAction: STREAMING_ACTION, waveform?: number[], audioLanguage?: SpeechToTextLanguage - ) => _model.streamingTranscribe(streamAction, waveform, audioLanguage), + ) => + controllerInstance.streamingTranscribe( + streamAction, + waveform, + audioLanguage + ), }; }; diff --git a/packages/react-native-executorch/src/hooks/useNonStaticModule.ts b/packages/react-native-executorch/src/hooks/useNonStaticModule.ts index bebd41d187..fe0e9860ca 100644 --- a/packages/react-native-executorch/src/hooks/useNonStaticModule.ts +++ b/packages/react-native-executorch/src/hooks/useNonStaticModule.ts @@ -29,7 +29,7 @@ export const useNonStaticModule = < const [isReady, setIsReady] = useState(false); const [isGenerating, setIsGenerating] = useState(false); const [downloadProgress, setDownloadProgress] = useState(0); - const _model = useMemo(() => new module(), [module]); + const moduleInstance = useMemo(() => new module(), [module]); useEffect(() => { if (preventLoad) return; @@ -39,7 +39,7 @@ export const useNonStaticModule = < setError(null); try { setIsReady(false); - await _model.load(model, setDownloadProgress); + await moduleInstance.load(model, setDownloadProgress); setIsReady(true); } catch (err) { setError((err as Error).message); @@ -47,18 +47,18 @@ export const useNonStaticModule = < })(); return () => { - _model.delete(); + moduleInstance.delete(); }; // eslint-disable-next-line react-hooks/exhaustive-deps - }, [_model, ...Object.values(model), preventLoad]); + }, [moduleInstance, ...Object.values(model), preventLoad]); const forward = async (...input: ForwardArgs): Promise => { if (!isReady) throw new Error(getError(ETError.ModuleNotLoaded)); if (isGenerating) throw new Error(getError(ETError.ModelGenerating)); try { setIsGenerating(true); - return await _model.forward(...input); + return await moduleInstance.forward(...input); } finally { setIsGenerating(false); } From 68b8e6ff81820998b13260c0613309e8a9c7596b Mon Sep 17 00:00:00 2001 From: jakmro Date: Tue, 5 Aug 2025 11:00:00 +0200 Subject: [PATCH 19/20] style: format code --- packages/react-native-executorch/src/constants/ocr/models.ts | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/packages/react-native-executorch/src/constants/ocr/models.ts b/packages/react-native-executorch/src/constants/ocr/models.ts index 18ce2500a3..f16ea5728c 100644 --- a/packages/react-native-executorch/src/constants/ocr/models.ts +++ b/packages/react-native-executorch/src/constants/ocr/models.ts @@ -10,7 +10,10 @@ const DETECTOR_CRAFT_320_MODEL = `${URL_PREFIX}-detector-craft/${VERSION_TAG}/xn type RecognizerSize = 64 | 128 | 256 | 512; -const createHFRecognizerDownloadUrl = (alphabet: keyof typeof alphabets, size: RecognizerSize) => +const createHFRecognizerDownloadUrl = ( + alphabet: keyof typeof alphabets, + size: RecognizerSize +) => `${URL_PREFIX}-recognizer-crnn.en/${VERSION_TAG}/xnnpack/${alphabet}/xnnpack_crnn_${alphabet}_${size}.pte`; const RECOGNIZER_ENGLISH_CRNN_512 = createHFRecognizerDownloadUrl( From e4c90f1af4c1dd614bf174d6ceabd13510936107 Mon Sep 17 00:00:00 2001 From: jakmro Date: Tue, 5 Aug 2025 11:04:52 +0200 Subject: [PATCH 20/20] docs: revert ExecutorchModule --- .../ExecutorchModule.md | 33 +++++++------------ 1 file changed, 11 insertions(+), 22 deletions(-) diff --git a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md index 958753291b..d7e9f2678a 100644 --- a/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md +++ b/docs/docs/03-typescript-api/03-executorch-bindings/ExecutorchModule.md @@ -16,25 +16,22 @@ import { const shape = [1, 3, 640, 640]; const input = new Float32Array(1 * 3 * 640 * 640); -// Creating an instance -const executorchModule = new ExecutorchModule(); - // Loading the model -await executorchModule.load(STYLE_TRANSFER_CANDY); +await ExecutorchModule.load(STYLE_TRANSFER_CANDY); // Running the model -const output = await executorchModule.forward(input, shape); +const output = await ExecutorchModule.forward(input, shape); ``` ### Methods -| Method | Type | Description | -| -------------------- | ------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -| `load` | `(modelSource: ResourceSource, onDownloadProgressCallback?: (progress: number) => void): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. | -| `forward` | `(input: ETInput, shape: number[]): Promise` | Executes the model's forward pass, where `input` is a JavaScript typed array and `shape` is an array of integers representing input Tensor shape. The output is a Tensor - raw result of inference. | -| `loadMethod` | `(methodName: string): Promise` | Loads resources specific to `methodName` into memory before execution. | -| `loadForward` | `(): Promise` | Loads resources specific to `forward` method into memory before execution. Uses `loadMethod` under the hood. | -| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. | +| Method | Type | Description | +| -------------------- | ------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | +| `load` | `(modelSource: ResourceSource): Promise` | Loads the model, where `modelSource` is a string that specifies the location of the model binary. | +| `forward` | `(input: ETInput, shape: number[]): Promise` | Executes the model's forward pass, where `input` is a JavaScript typed array and `shape` is an array of integers representing input Tensor shape. The output is a Tensor - raw result of inference. | +| `loadMethod` | `(methodName: string): Promise` | Loads resources specific to `methodName` into memory before execution. | +| `loadForward` | `(): Promise` | Loads resources specific to `forward` method into memory before execution. Uses `loadMethod` under the hood. | +| `onDownloadProgress` | `(callback: (downloadProgress: number) => void): any` | Subscribe to the download progress event. |
    Type definitions @@ -42,7 +39,7 @@ const output = await executorchModule.forward(input, shape); ```typescript type ResourceSource = string | number | object; -type ETInput = +export type ETInput = | Int8Array | Int32Array | BigInt64Array @@ -54,15 +51,7 @@ type ETInput = ## Loading the model -To load the model, use the `load` method. It accepts the following arguments: - -**`modelSource`** - A string that specifies the location of the model binary. - -**`onDownloadProgressCallback`** - (Optional) Function called on download progress. - -This method returns a promise, which can resolve to an error or void. - -For more information on loading resources, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. +To load the model, use the `load` method. It accepts the `modelSource` which is a string that specifies the location of the model binary. For more information, take a look at [loading models](../../01-fundamentals/02-loading-models.md) page. This method returns a promise, which can resolve to an error or void. ## Running the model