Unable to load and run model using vLLM
#5
by
hailexiao
- opened
When I try to load and run the model using vLLM with the command
vllm serve "mohbattharani/LLaVA-Chef"
I receive the following error:
TypeError: Special token pad_token has to be either str or AddedToken but got: <class 'int'>
I noticed that in special_tokens_map.json, the value for the key "pad_token" is -100; should it be "-100" instead, since a string is expected?