[NVBUG-6379624][fix] Enable W4A8 checkpoint loading for Gemma4 K=V layers#16797

Hudayday merged 1 commit into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

Hudayday:tianruih/nvbug-6379624-gemma4-w4a8-kv-scaleHudayday/TensorRT-LLM:tianruih/nvbug-6379624-gemma4-w4a8-kv-scaleCopy head branch name to clipboard