Transformers: Can't load config for [community model]

[abedkhooli]

Although I can use a fine-tuned GPT2 model from code, the model page complains about the config file (which is already uploaded).
at https://huggingface.co/akhooli/gpt2-small-arabic-poetry (for a prompt), I get:

Can't load config for 'akhooli/gpt2-small-arabic-poetry'. Make sure that: - 'akhooli/gpt2-small-arabic-poetry' is a correct model identifier listed on 'https://huggingface.co/models' - or 'akhooli/gpt2-small-arabic-poetry' is the correct path to a directory containing a config.json file 

[julien-c]

Currently working on a fix. Will update here

[julien-c]

Should be fixed now

[rohanrajpal]

Hey I'm facing the same issue with two models I uploaded today
https://huggingface.co/rohanrajpal/bert-base-en-hi-codemix-cased?text=I+like+you.+I+love+you
https://huggingface.co/rohanrajpal/bert-base-en-es-codemix-cased?text=I+like+you.+I+love+you

Here are the config files
https://s3.amazonaws.com/models.huggingface.co/bert/rohanrajpal/bert-base-en-es-codemix-cased/config.json
https://s3.amazonaws.com/models.huggingface.co/bert/rohanrajpal/bert-base-en-hi-codemix-cased/config.json

[julien-c]

pinging @mfuntowicz on this

[abedkhooli]

I am seeing the same issue with a new model I just uploaded (akhooli/xlm-r-large-arabic-sent). It works if called in code through HF pipeline.

[donal1]

Seeing the same issue, pinging @mfuntowicz, @julien-c

https://huggingface.co/donal/Pro_Berta?text=The+goal+of+life+is+%3Cmask%3E.

All the files seem to be in the right place.

Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/merges.txt
Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/special_tokens_map.json
Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/training_args.bin
Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/pytorch_model.bin
Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/config.json
Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/tokenizer_config.json
Your file now lives at:
https://s3.amazonaws.com/models.huggingface.co/bert/donal/Pro_Berta/vocab.json

[abedkhooli]

You probably are already aware of this, but inference worked for a bit and broke again (Can't load config for [model]).

[mfuntowicz]

I had to revert the change I did because it was breaking one workflow we have on api-inference side. I'm working on having a stable patch by today, sorry for the inconvenience

[mfuntowicz]

Should be fixed now. Let us know 👍

[donal1]

Yup works

[abedkhooli]

Confirmed (existing model then updated). Thanks for the fix and for HF great work!

[rohanrajpal]

It works now. Thanks! @mfuntowicz