Naman Jain

Hi, thanks a lot for the PR. It seems the COCO/OKS codebase has changed a lot since I wrote this code. Unfortunately, I am no longer maintaining the repo and...

I was not sure if this should be a separate issue but I found that pruning does not work unless we explicitly add `training=True` argument in the model call. (This...

Given any span you can use the function to get a list of labels ```py def get_span_labels(span: str) -> List[str]: labels = span._.labels if len(labels) == 0: doc = span.doc...

Sorry for hijacking this but I am getting an OOM error on using 8 A100 40GB GPUs. Since the total GPU memory is more than your setup I am curious...

Thank you very much for the quick response! I used `run_clm.py` with `--gradient_checkpointing True` and it seems to work for GPTJ with `block_size` 1024! However, it does not work for...

Created a post at https://discuss.huggingface.co/t/fsdp-oom-issue-and-comparision-to-deepspeed/44292/1