Sylvain Gugger

Thanks for flagging! The PR linked above should solve this.

I think one issue with the list of TODOs is fine.

Thanks again for your contribution!

Do you have a full example of the error you are reporting I could run? I am unable to reproduce it. Something like the [test](https://github.com/huggingface/transformers/blob/ab2006e3d6db88654526a4169e65d4bfc52da2e3/tests/models/auto/test_tokenization_auto.py#L234) of this feature we could...

`huggingface_hub` does not take into account redirections in its download methods. The issue was given low priority from what I understand, you can bug folks internally to show it's a...

I don't think it makes sense to create an auto-map just for this model, and the pipeline can be done as a [custom pipeline with code on the Hub](https://huggingface.co/docs/transformers/add_new_pipeline#share-your-pipeline-on-the-hub). If/when...

Are you sure you pasted the exact command you ran? I have no error when trying it on my side and the config is successfully updated. To use distributed training,...

@datquocnguyen We can't merge anything that has any breaking change on the existing tokenizer, as I said before.

Hi @datquocnguyen. It's amazing that you added those two new fast tokenizers. However we need PRs to be focused on one thing. Would you terribly mind splitting it in three...

You need to install `black==22.3` to have the same results as the CI.