api-inference is not working

#6
by baby1 - opened

curl https://api-inference.huggingface.co/models/chavinlo/alpaca-native
-X POST
-d '{"inputs": "Can you please let us know more details about your "}'
-H "Authorization: Bearer xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx replace with my own token

response is :
{"error":"The model_type 'llama' is not recognized."}

It seem transformers source has not been updated from the lastest PR ?

either that or I need to change LLaMa to llama

tokenizer_class in tokenizer_config.json needs to be changed to: LlamaTokenizer

The model_type 'llama' is not recognized. It could be a bleeding edge model, or incorrect

Is it normal for the model to sometimes just to parrot the prompt with no additional text? Or is there a trick to setting the temperature, num_beams, top_p, etc.? I find it's hit and miss.

Sign up or log in to comment