Are there any plans to make BF16/FP8 AWQ INT4 version of Qwen/Qwen3.5-397B-A17B?
#7
by zuuky - opened
The current MiniMax-M2.5-BF16-INT4-AWQ is the best version I've ever used, no one, awesome !!!
Qwen/Qwen3.5-397B-A17B naturally supports multimodality, and the effect of multimodality looks good
The Qwen team has released a INT4 GPTQ version but it's already just 236GB of weights which is beyond my 2x RTX Pro 6000 setup so I wouldn't be able to test my quant. :/