VLLM contribution

[arunpatala]

Thanks for this great repo. I would like to run EAGLE with VLLM.
I would like to contribute to the VLLM implementation. If you are
already working on a branch I would like to help. If not, if you
can point me what changes need to be made. That would be most
helpful.

[yuhuili]

Very welcome! The draft model of EAGLE is essentially a Decoder layer of LLaMA, so two models should be created using the backend of vLLM, and the base model should be modified to output hidden states. The hidden states and input_ids should be concatenated, and then input into the draft model after dimension reduction. The most important thing to note is to modify the block table of PagedAttention to truncate rejected tokens.

[arunpatala]

Thanks for the information. I will let you know how it goes.