Shengliang Xu

shengliangx

nvidia

·

AI & ML interests

LLM

Recent Activity

new activity 2 days ago

nvidia/GLM-5.3-Flash-NVFP4:MTP layer is unquantized but missing from exclude_modules, so vLLM cannot load it for speculative decoding

new activity 3 days ago

nvidia/GLM-5.3-Flash-NVFP4:Why don't the commits in this repo have the "verified" badge?

new activity 3 days ago

nvidia/GLM-5.3-Flash-NVFP4:MTP works after adding layer 45 to the quantization exclusion lists

View all activity

Organizations

shengliangx 's datasets

None public yet