Cannot use the model with vllm

#23
by NahuelCosta - opened

Using the command provided here:

# Load and run the model:
vllm serve "liuhaotian/llava-v1.5-7b"

I get this error (both for the 7b and 13b models):

"LlavaLlamaForCausalLM has no vLLM implementation and the Transformers implementation is not compatible with vLLM."

Is there any way to fix it or get it working?

Thanks in advance.

Sign up or log in to comment