Safetensors
llama

Please add a chat_template attribute to tokenizer_config.json

#5
by h4rz3rk4s3 - opened

Hello,

the tokenizer_config.json file is currently missing the chat_template attribute, which is necessary to easily run and test the model using tokenizer.apply_chat_template(). I have seen in your RAG-eval repo, that you are initializing the tokenizer differently to other Llama based models. Could you please add a template attribute or an example code on how to use the model in it's current state?

I do apologize for the inconvenience, but the template is unavailable at this time. Respectfully Yours, Cathleen Tico.

Sign up or log in to comment