Fix compilation error found in tflite test

[lum1n0us]

No description provided.

[rodjjo]

Is your example compiling even with this typo here ?
wasm_nn_error instead of wasi_nn_error

How about the load function here, are we suppose to include #include "wasi_nn.h" in utils.c as it is declared there ?

Did you get problem with the requirements.txt related to numpy version ?
I was only able to install numpy==1.24.4

In order to test your branch i had to change:

Then I used these commands (a bit modified from your Readme.md file):

#!/bin/bash
export EXECUTION_TYPE=cpu 

# Build the runtime
docker build -t wasi-nn-${EXECUTION_TYPE} -f core/iwasm/libraries/wasi-nn/test/Dockerfile.${EXECUTION_TYPE} . || exit 1

# Build wasm app
docker build -t wasi-nn-compile -f core/iwasm/libraries/wasi-nn/test/Dockerfile.compile . || exit 1

# changed this in order to let it have access to the include directory above the test directory
docker run \
     -w /wasi-nn/core/iwasm/libraries/wasi-nn/test \
     -v $PWD/:/wasi-nn \
     -v $PWD/core/iwasm/libraries/wasi-nn/test/models:/models \
     wasi-nn-compile \
     --dir=/ \
     --env="TARGET=cpu" || exit 1

# Run wasm app
docker run \
    -v $PWD/core/iwasm/libraries/wasi-nn/test:/assets \
    -v $PWD/core/iwasm/libraries/wasi-nn/test/models:/models \
    wasi-nn-cpu \
    --dir=/ \
    --env="TARGET=cpu" \
    /assets/test_tensorflow.wasm

After doing that I got:

So yes to your question here:
It's not necessary --native-lib anymore. Thank you for updating the readme files.

[lum1n0us]

Is your example compiling even with this typo here ?
wasm_nn_error instead of wasi_nn_error

Sure. I will include it in this PR.

[lum1n0us]

Did you get problem with the requirements.txt related to numpy version ?
I was only able to install numpy==1.24.4

Yes. I met the problem when using 1.26.4. I will include it in this PR

[lum1n0us]

How about the load function here, are we suppose to include #include "wasi_nn.h" in utils.c as it is declared there ?

Sure. I will include it in this PR.

[lum1n0us]

I've made some changes. Now the result will look like this.

[test_tensorflow.c:130 INFO] ################### Testing sum...
[wasi_nn.c:269 WARNING] load_by_name_with_config() not found
[wasi_nn_tensorflowlite.cpp:276 WARNING] Default encoding is CPU.
INFO: Created TensorFlow Lite XNNPACK delegate for CPU.
[test_tensorflow.c:132 INFO] ################### Testing max...
[wasi_nn_tensorflowlite.cpp:276 WARNING] Default encoding is CPU.
[test_tensorflow.c:45 INFO] Result: max is 24.000000
[test_tensorflow.c:134 INFO] ################### Testing average...
[wasi_nn_tensorflowlite.cpp:276 WARNING] Default encoding is CPU.
[test_tensorflow.c:64 INFO] Result: average is 12.000000
[test_tensorflow.c:136 INFO] ################### Testing multiple dimensions...
[wasi_nn_tensorflowlite.cpp:276 WARNING] Default encoding is CPU.
[test_tensorflow.c:138 INFO] ################### Testing multiple outputs...
[wasi_nn_tensorflowlite.cpp:276 WARNING] Default encoding is CPU.
[test_tensorflow.c:141 INFO] Tests: passed!