vllm
vllm copied to clipboard
[VLM] Support multimodal inputs for Florence-2 models
FIX #5934