regisss

Results 73 comments of regisss

Replace `-m torch.distributed.run` by `torchrun`

However `torchrun` has only been available since the release of `torch` 1.10. I guess we want to keep compatibility with some previous versions of `torch` right? @sgugger @ArthurZucker

Repo card metadata block was not found. Setting CardData to empty

Where have you downloaded the zip files? It's likely that they are not in a folder called `data` in the current directory.

Support mixtral long sequence 32k with bs 2

@jychen-habana Is this PR different from #903 ?

Update to transformers 4.39.1+

Transformers v4.40 should be released soon so I'll work on directly jumping from 4.38 to 4.40

Update to transformers 4.39.1+

I'll push a branch compatible with Transformers v4.40 next week. If there is no major regression, it should be merged a few days after.

Facing issue while running the git repository for text generation

@VenkateshPasumarti Can you share the command you executed and the whole error message you got please?

run_lora_clm.py support for other datasets

Yes, I agree that would be much better. Can you share a command line that fails? And the error message you get please?

what the difference between decoder_model.onnx and decoder_with_past_model.onnx

@ayansengupta17 It has been moved [here](https://huggingface.co/docs/optimum/onnxruntime/usage_guides/models#export-and-inference-of-sequencetosequence-models).

Converted LayoutLM ONNX model - Required input `bbox` missing from input feed `['input_ids', 'attention_mask', 'token_type_ids']`

This is an Optimum issue and not Transformers. Can you show the whole error message so that we can see which missing output triggerred the error please?

Workarounds for failures in using HF Transformers auto classes in Classification sample

@sandeep-maddipatla I think this change should first be merged into Transformers and then we can add it here. Would you like to open a PR in Transformers?