lukealonso
/

MiniMax-M2.5-NVFP4

8-bit precision

Model card Files Files and versions

Resources

View closed (0)

VLLM error for kv weight scaling - workaround

#6 opened about 10 hours ago by

Thanks for your effort

#5 opened about 17 hours ago by

fp8 kv cache

#4 opened 1 day ago by

KeyError: '110.w1.input_scale' with TRT

#3 opened 2 days ago by

"w1_weight_scale_2 must match w3_weight_scale_2. Accuracy may be affected."

#2 opened 3 days ago by

Here's the vLLM recipe I'm using with 2x RTX Pro 6000

#1 opened 3 days ago by