Upload onnx/language_model_q4f16.onnx with huggingface_hub

Upload onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload onnx/speech_encoder_q4f16.onnx with huggingface_hub

Upload onnx/embed_tokens_q4f16.onnx with huggingface_hub

Upload onnx/language_model_q4f16.onnx with huggingface_hub

Upload onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload onnx/speech_encoder_q4f16.onnx with huggingface_hub

Upload onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload onnx/speech_encoder_q4f16.onnx with huggingface_hub

Upload onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload onnx/speech_encoder_q4f16.onnx with huggingface_hub

Upload onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload onnx/embed_tokens_q4f16.onnx with huggingface_hub

Upload onnx/language_model_q4f16.onnx with huggingface_hub

Upload onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload onnx/speech_encoder_q4f16.onnx with huggingface_hub

fix: use official ONNX models with Cast nodes for Slice (preserves INT64 for Unsqueeze)

fix: use official ONNX models with Cast nodes for Slice (preserves INT64 for Unsqueeze)

fix: use official ONNX models with Cast nodes for Slice (preserves INT64 for Unsqueeze)

fix: use official ONNX models with Cast nodes for Slice (preserves INT64 for Unsqueeze)

fix: insert Cast nodes for Slice int64 inputs (preserves Unsqueeze int64)

fix: insert Cast nodes for Slice int64 inputs (preserves Unsqueeze int64)

fix: insert Cast nodes for Slice int64 inputs (preserves Unsqueeze int64)

fix: insert Cast nodes for Slice int64 inputs (preserves Unsqueeze int64)

fix: selective int64->int32 for Slice only (keep Unsqueeze int64)

fix: selective int64->int32 for Slice only (keep Unsqueeze int64)

fix: selective int64->int32 for Slice only (keep Unsqueeze int64)

fix: selective int64->int32 for Slice only (keep Unsqueeze int64)

fix: convert int64 to int32 for WebGPU compatibility

fix: convert int64 to int32 for WebGPU compatibility

fix: convert int64 sentinels to int32 for WebGPU compatibility

fix: convert int64 sentinels to int32 for WebGPU compatibility

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Delete onnx/speech_encoder.onnx_data with huggingface_hub

Delete onnx/speech_encoder.onnx with huggingface_hub

Delete onnx/embed_tokens.onnx_data with huggingface_hub

Delete onnx/embed_tokens.onnx with huggingface_hub

Delete onnx/conditional_decoder.onnx_data with huggingface_hub

Delete onnx/conditional_decoder.onnx with huggingface_hub

Delete onnx/speech_encoder_q4f16.onnx_data with huggingface_hub

Delete onnx/speech_encoder_q4f16.onnx with huggingface_hub

Delete onnx/embed_tokens_q4f16.onnx_data with huggingface_hub

Delete onnx/embed_tokens_q4f16.onnx with huggingface_hub

Delete onnx/conditional_decoder_q4f16.onnx_data with huggingface_hub

Delete onnx/conditional_decoder_q4f16.onnx with huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

Upload folder using huggingface_hub

initial commit