onnx conversion for custom trained trocr base stage1

[Mir-Umar]

Feature request

I have trained the base stage1 trocr on my custom dataset having multiline images. The trained model gives good results while using the default torch format for loading the model. But while converting the model to onnx, the model detects only first line or part of it in first line. I have used this https://github.com/huggingface/transformers/issues/19811#issuecomment-1303072202 for converting the model to onnx. Can you kindly provide the insights about what i should do differently, in order to get the desired multiline output from the onnx converted model.

Motivation

How to update the onnx conversion of trocr in order to support multiline trocr trained model

Your contribution

Trained a trocr base stage1 model for multiline dataset.

[fxmarty]

Hi @Mir-Umar, thank you for the report. Could you provide a reproducible example of the issue? Thank you!