ONNX version of open-ai/whisper large model

[Prashanth-dvs]

I have exported openai/whisper model to ONNX but size of ONNX file is about 4.5GB. After quantization size reduces to 1.3GB. Kindly suggest me some method to reduce the onnx size further. Also while performing optimization getting run error

KeyError: "ONNX Runtime doesn't support the graph optimization of whisper yet. Only ['bert', 'gpt2', 'bart'] are supported. If you want to support whisper please propose a PR or open up an issue in ONNX Runtime:https://github.com/microsoft/onnxruntime."

Please help anyone