Not able to used qlora models with vllm

[royrajjyoti1]

I have trained falcon 7b model with qlora but the inference time for outputs is too high.So I want to use vllm for increasing the inference time for that I have used a code snippet to load the model path
llm = LLM(model="/content/trained-model/").
But I am getting an error :

OSError: /content/trained-model/ does not appear to have a file named config.json. Checkout 
'https://huggingface.co//content/trained-model//None' for available files.

[royrajjyoti1]

Thank you @zhuohan123 for the reply.
Can you provide me the ETD for falcon model because I have checked 4 days before it will be up in few days.( [https://github.com//issues/195])

[ehartford]

You just need to merge the model. Vllm doesn't support LoRA.

[fabianlim]

@ehartford but merging has to be done in higher precision. Doesnt that defeat the purpose of wanting to have the base weights in low precision to speed up inference?

[hmellor]

Closing in favour of the feature request #3225