vllm
[Core] Refactor `QKVCrossParallelLinear` implementation to support BNB 4-bit quantization
#14545
Merged

Loading