Hugging Face's logo

LibertAIDAI
/
GLM-5.3-Flash-NVFP4

Image-Text-to-Text
Safetensors
English
Chinese
glm5_next
nvfp4
blackwell
sglang
vllm
glm
glm-5
Mixture of Experts
multimodal
modelopt
conversational
8-bit precision
Model card Files
xet
 Community
10

The model's single-inference speed measured on two DGX systems is only 16 tokens/s, which does not meet productivity requirements.

#9
by SAJDOQPDPS - opened
Discussion
SAJDOQPDPS

https://github.com/beastllama/GLM-5.3-Flash-DFlash2-SGLang-2x-DGX-Spark
jon1012
 LibertAI org

You can try this recipe: https://github.com/Libertai/glm53-flash-vllm-gb10

Sign up or log in to comment