Read all the patches in and absorb then into torch_lib if applicable. For example we can update TopK to specify the sorted attribute explicitly. etc. https://github.com/huggingface/transformers/blob/98445c3c81483c076fd1e763adb8b4877d98c242/src/transformers/exporters/exporter_onnx.py#L647
Read all the patches in and absorb then into torch_lib if applicable. For example we can update TopK to specify the sorted attribute explicitly. etc.
https://github.com/huggingface/transformers/blob/98445c3c81483c076fd1e763adb8b4877d98c242/src/transformers/exporters/exporter_onnx.py#L647