ERROR:hf-to-gguf:Model BertForTokenClassification is not supported

#4
by Raphy10-Collab - opened

Trying to convert this PyTorch model https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-italian to a Gguf model, but got ERROR:hf-to-gguf:Model BertForTokenClassification is not supported

(.venv) raphy@raohy:~/llama.cpp$ python3 ./convert_hf_to_gguf.py bert-italian-finetuned-ner/ --outtype f16 --outfile ./models/bert-italian-finetuned-ner.gguf
INFO:hf-to-gguf:Loading model: bert-italian-finetuned-ner
ERROR:hf-to-gguf:Model BertForTokenClassification is not supported

How to make it work?

Hi @Raphy10-Collab ,
GGUF is for generative models, BERT is an encoder, isn't supported.
Check this discussion: https://huggingface.co/google-bert/bert-base-uncased/discussions/66

Thank you @nickprock for clarifying me this import aspect

Raphy10-Collab changed discussion status to closed

Sign up or log in to comment