nb-whisper-tiny / export_models.sh
pere's picture
Update export_models.sh
00fb3d5
raw
history blame
2.07 kB
#!/bin/bash
pip install "optimum[exporters]>=1.14.1" tensorflow
git lfs track *.onnx*
python << END
from transformers import WhisperForConditionalGeneration, TFWhisperForConditionalGeneration, WhisperTokenizerFast
import shutil
# Backup generation_config.json
shutil.copyfile('./generation_config.json', './generation_config_backup.json')
print("Saving model to PyTorch...", end=" ")
model = WhisperForConditionalGeneration.from_pretrained("./", from_flax=True)
model.save_pretrained("./", safe_serialization=True)
model.save_pretrained("./")
print("Done.")
print("Saving model to TensorFlow...", end=" ")
tf_model = TFWhisperForConditionalGeneration.from_pretrained("./", from_pt=True)
tf_model.save_pretrained("./")
print("Done.")
# Restore the backup of generation_config.json
shutil.move('./generation_config_backup.json', './generation_config.json')
print("Saving model to ONNX...", end=" ")
from optimum.onnxruntime import ORTModelForSpeechSeq2Seq
ort_model = ORTModelForSpeechSeq2Seq.from_pretrained("./", export=True)
ort_model.save_pretrained("./onnx")
print("Done")
print("Saving model to CTranslate...", end=" ")
ct2-transformers-converter --model . --output_dir ct2
cp ct2/model.bin .
cp ct2/vocabulary.json .
cp config.json config_hf.json
jq -s '.[0] * .[1]' ct2/config.json config_hf.json > config.json
print("Done")
END
echo "Saving model to GGML (whisper.cpp)..."
wget -O convert-h5-to-ggml.py "https://raw.githubusercontent.com/ggerganov/whisper.cpp/94aa56f19eed8b2419bc5ede6b7fda85d5ca59be/models/convert-h5-to-ggml.py"
mkdir -p whisper/assets
wget -O whisper/assets/mel_filters.npz "https://github.com/openai/whisper/raw/c5d42560760a05584c1c79546a098287e5a771eb/whisper/assets/mel_filters.npz"
python ./convert-h5-to-ggml.py ./ ./ ./
rm ./convert-h5-to-ggml.py
rm -rf ./whisper
echo "Done"
echo "Quantizing GGML model..."
git clone --depth 1 https://github.com/ggerganov/whisper.cpp --branch v1.5.1
cd whisper.cpp/
make -j 32
make quantize -j 32
./quantize ../ggml-model.bin ../ggml-model-q5_0.bin q5_0
cd ..
rm -rf whisper.cpp
echo "Done"