Cannot export TrOCR to ONNX with past

[mht-sharma]

System Info

- `optimum` version: 1.6.4.dev0
- `transformers` version: 4.26.0
- Platform: Linux-5.4.0-125-generic-x86_64-with-glibc2.17
- Python version: 3.8.15
- Huggingface_hub version: 0.11.0
- PyTorch version (GPU?): 1.13.1+cu117 (cuda availabe: True)
- Tensorflow version (GPU?): 2.10.1 (cuda availabe: False)

Who can help?

No response

Information

• [ ] The official example scripts
• [ ] My own modified scripts

Tasks

• [ ] An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• [ ] My own task or dataset (give details below)

Reproduction

python -m optimum.exporters.onnx --model="microsoft/trocr-small-handwritten" --task=vision2seq-lm-with-past model_trocr_base   --for-ort

Error:

ValueError: Exporting past key values is not supported with TrOCR model!

Expected behavior

Exporting TrOCR to ONNX with past should work.

[fxmarty]

Hi, this is fixed in https://github.com/huggingface/optimum/pull/1456

[goodevile]

I got it with:

!optimum-cli export onnx --task image-to-text --model microsoft/trocr-small-printed trocr-small-printed_onnx/

[CrasCris]

I got it with:

!optimum-cli export onnx --task image-to-text --model microsoft/trocr-small-printed trocr-small-printed_onnx/

Thats the Task of TrOCR?, i dont find any docs for that, how i can load then to use it only for inferention ?