Open gbea opened 2 years ago
Hi, I'm currently adding support for encoder-decoder models to the Core ML exporter.
In theory this should work already but there are some issues with coremltools that prevent these kinds of models from being converted correctly (or at all).
I'm actively working on getting this fixed because we do want to support Core ML versions of these models.
Recently onnx export for AutoModelForVision2Seq has been added: https://github.com/huggingface/transformers/pull/19254 I would be really interested in converting this topology to coreml. Is this feature planned for coreml as well?
Thanks!