AttributeError: 'STLlamaConfig' object has no attribute 'lin_hidden_size'

[agoodge]

When I execute run_urbangpt.py with --model-name facebook/opt350m, I get the following error:

AttributeError: 'STLlamaConfig' object has no attribute 'lin_hidden_size'

All of the other arguments are left unchanged from the .sh file.

Can you tell me what is wrong? Thanks

[zxc2012]

I think the author mentioned the updated config file What I did was copying the content and pasting into my own huggingface "models--lmsys--vicuna-7b-v1.5-16k" folder

[LZH-YS1998]

If you have already downloaded all the files of the urbangpt model, can you run urbangpt_eval.sh normally? (sh urbangpt_eval.sh)