There are some applications such as Kaggle which require running without an internet connection.
At this moment the package can be downloaded along with the checkpoints, but still, the creation requires pulling details from the HF hub, so it'll be cool if we download the details offline to be able to use them instead of the online source...
I guess that the solution would be exposing this argument in Module init:
|
pretrained_model_name_or_path=None, |
There are some applications such as Kaggle which require running without an internet connection.
At this moment the package can be downloaded along with the checkpoints, but still, the creation requires pulling details from the HF hub, so it'll be cool if we download the details offline to be able to use them instead of the online source...
I guess that the solution would be exposing this argument in Module init:
detoxify/detoxify/detoxify.py
Line 20 in 0cccd59