card model contains a non-working example of use

#1
by vlinkworkacc - opened

My pipeline stopped working after I tried to load the XProvence model. This line of code seems to be the culprit:
model = AutoModel.from_pretrained("naver/xprovence", trust_remote_code=True)
The pipeline was working perfectly until yesterday. Has anyone encountered a similar issue or know what might have changed?

vlinkworkacc changed discussion title from 401 Client Error. to card model contains a non-working example of use
NAVER LABS Europe org

Hi @vlinkworkacc !

Thanks for pointing us to this bug in the readme! I changed it, it should be xprovence = AutoModel.from_pretrained("naver/xprovence-reranker-bgem3-v2", trust_remote_code=True)

It's like this now: AttributeError: 'XLMRobertaModel' object has no attribute 'process'. :)
it looks like the problem is in config.json, maybe missing: "auto_map": { "AutoModel": "modeling_xprovence_hf.XProvence"}

Hi @vlinkworkacc !

Thanks for pointing us to this bug in the readme! I changed it, it should be xprovence = AutoModel.from_pretrained("naver/xprovence-reranker-bgem3-v2", trust_remote_code=True)

NAVER LABS Europe org

Hello @vlinkworkacc
It was indeed an issue with the config.json.
We have fixed it and the model should be working properly now.

Now it really works as it should. Thank you so much for your help.

vlinkworkacc changed discussion status to closed

Sign up or log in to comment