Jordan Parker

@cjdsellers yeah a while loop is resolving the pagination issues

I'm a bit stuck with some of these test cases I have written. It looks like a float point error is present and some issues parsing the c types. Any...

Ahh that great! Thank you.

For those interested, I found that the HF implementation is set up for Gradient Accumulation. Enable it with: ```py self.model = YolosForObjectDetection.from_pretrained( self.hparams.pretrained_model_name_or_path, config=config, ignore_mismatched_sizes=True ) self.model.gradient_checkpointing_enable() ``` I was...

@logan-markewich I like your approach. What's your thoughts on replacing the BART encoder with one of the long-rage Transformer architecture (e.g. LongFormer or Performer)? I don't have much experience with...

@BendaCoding are the types committed and available through npm?

I am getting the same error. It might be a cuda / pytorch version issue. https://github.com/facebookresearch/detectron2/issues/2837

Yeah dataset is in COCO format. I also tested the training script with CPU only. That worked fine so I think its a CUDA / Pytorch versioning error like https://github.com/facebookresearch/detectron2/issues/2837....

Sounds like you should look at the transformers implementation of LayoutLMv3. https://huggingface.co/docs/transformers/main/en/model_doc/layoutlmv3#transformers.LayoutLMv3ForTokenClassification When doing token classification it will predict a label for each token based on the token and the...

> Currently, LayoutLMv3 in Transformers does not support object detection ([see @NielsRogge's reply below](https://github.com/huggingface/transformers/pull/17060#issuecomment-1132626756)). > > > unfortunately I'm (for now) not planning to add the object detection part, because...