Ehsan M. Kermani

Thanks for creating this feature request! It's planned but we don't have a fixed timeline yet. Stay tuned!

Thanks for reporting! what version of `max -v` are you using?

Thank you! could you share more about what type of model is having this issue and perhaps provide a reproducing example?

Thanks for reporting! that's odd and seems a huggingface issue. What version of `transformers` are you using?

Please check out [this explainer](https://github.com/modularml/max/issues/79#issuecomment-1974150979)

I'm getting ``` Traceback (most recent call last): File "/Users/ehsan/workspace/transformers/utils/check_task_guides.py", line 58, in "asr.mdx": transformers_module.models.auto.modeling_auto.MODEL_FOR_CTC_MAPPING_NAMES, File "/Users/ehsan/workspace/transformers/src/transformers/utils/import_utils.py", line 1150, in __getattr__ raise AttributeError(f"module {self.__name__} has no attribute {name}") AttributeError: module...

As long as the operators are supported by MXNet ONNX, there should be no problem. Checkout this [tutorial](https://mxnet.apache.org/api/python/docs/tutorials/deploy/export/onnx.html) to start.

I've moved on from the changes to this project. But in general, you'd want to export the underlying pytorch model to onnx. Afaik the MXNet part is pretty much inactive...

Curious have you resolved the issue?