clechristophe comments

Repositories
Issues
Comments

Results 2 comments of


                                            clechristophe

trafficstars

Integrated Gradients with pre-trained CamembertForSequenceClassification model explainer

Hi @jklaise, Yes the example notebook works fine with transformers 4.16.2. When using the wrapped version, I've got the following error when computing the predictions: ``` --------------------------------------------------------------------------- RecursionError Traceback (most...

CUDA OOM during ckpt saving for Llama2-70b

I have the same problem. I think everything is brought back to the first rank before saving and it causes CUDA OOM. I was able to save one .distcp file...