[None][fix] Fix Qwen3.5 MoE fallback, GDN alignment, FP8 activation, and draft KV cache - #17120

[None][fix] Fix Qwen3.5 MoE fallback, GDN alignment, FP8 activation, and draft KV cache#17120

Wanli-Jiang merged 4 commits into

NVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

Wanli-Jiang:user/williamj/qwen35-bug-fix3Wanli-Jiang/TensorRT-LLM:user/williamj/qwen35-bug-fix3Copy head branch name to clipboard