fine-tuning table transformer model

[skwskwskwskw]

Hi,

Is there a tutorial on how to fine-tune table transformer (table detaction model) with custom dataset?

And how should the annotated data looks like?

Thanks,

[NielsRogge]

Hi,

Table Transformer follows the same architecture as DETR, hence you can use this notebook for fine-tuning: https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb.

Make sure to replace the image processor and model accordingly.