Training Dataset Tags - Not Showing #624
Description
This is for bugs only
Did you already ask in the discord?
No (Don't use Discord)
You verified that this is a bug and not a feature request or question by asking in the discord?
No Same as above
Describe the bug
After a successful training run (models work correctly) in the info pane there Training Dataset Tags are not present.
I have tried adding kohyaSS scripts/code to force it to show Training Dataset Tags, like:
save_metadata: true
save_caption_metadata: true
save_dataset_metadata: true
store_dataset_info: true
write_tag_frequency: true
Nothing seems to work, is this not a feature in AI-Toolkit? Or am I just missing something so simple and obvious? Any help would be appreciated. Thank you!~
Example of what it should look like:
What it actually looks like:
My Configuration for Training:
config:
name: "Charact3r (illust) v2b"
process:
- type: "diffusion_trainer"
training_folder: "E:\\PC BACKUP\\ai-toolkit\\output"
sqlite_db_path: "./aitk_db.db"
device: "cuda"
trigger_word: charact3r
performance_log_every: 10
network:
type: "lora"
linear: 24
linear_alpha: 24
conv: 12
conv_alpha: 12
lokr_full_rank: true
lokr_factor: -1
network_kwargs:
ignore_if_contains: []
save:
dtype: "bf16"
save_every: 450
max_step_saves_to_keep: 4
save_format: "safetensors"
save_metadata: true
push_to_hub: false
datasets:
- folder_path: "E:\\PC BACKUP\\ai-toolkit\\datasets/charact3r"
mask_path: null
mask_min_value: 0.1
default_caption: ""
caption_ext: "txt"
caption_dropout_rate: 0.01
cache_latents_to_disk: true
is_reg: false
network_weight: 1
resolution:
- 1024
controls: []
shrink_video_to_frames: true
num_frames: 1
do_i2v: false
flip_x: false
flip_y: false
train:
batch_size: 1
bypass_guidance_embedding: false
steps: 450
gradient_accumulation: 1
train_unet: true
train_text_encoder: true
gradient_checkpointing: true
noise_scheduler: "ddpm"
optimizer: "adamw8bit"
timestep_type: "sigmoid"
content_or_style: "balanced"
optimizer_params:
weight_decay: 0.04
unload_text_encoder: false
cache_text_embeddings: false
lr: 0.0001
ema_config:
use_ema: false
ema_decay: 0.99
skip_first_sample: false
force_first_sample: false
disable_sampling: true
dtype: "bf16"
diff_output_preservation: false
diff_output_preservation_multiplier: 1
diff_output_preservation_class: "person"
switch_boundary_every: 1
loss_type: "mse"
ema_config?:
ema_decay: 0.99
logging:
log_every: 1
use_ui_logger: true
model:
name_or_path: "E:\\PC BACKUP\\ai-toolkit\\base_models\\illustriousXL_v01.safetensors"
quantize: false
qtype: "qfloat8"
quantize_te: false
qtype_te: "qfloat8"
arch: "sdxl"
low_vram: false
model_kwargs: {}
sample:
sampler: "ddpm"
sample_every: 100
width: 1024
height: 1024
samples: []
neg: ""
seed: 42
walk_seed: true
guidance_scale: 5.5
sample_steps: 25
num_frames: 1
fps: 1
meta:
name: "Trained Model Test"
version: "1.0"```