vllm
[Transformers backend] Support LoRA on lm_head and on any input embedding name
#66413
Merged

Loading