amd
/

Mixtral-8x7B-Instruct-v0.1-WMXFP4FP8-AMXFP4FP8-AMP-KVFP8

Model card Files Files and versions

Mixtral-8x7B-Instruct-v0.1-WMXFP4FP8-AMXFP4FP8-AMP-KVFP8 / tokenizer.json

XuebinWang's picture

update the AMP model with better accuracy numbers and lower effective bitwidth (5.6) (#2)

1d1cfda verified about 1 month ago

history contribute delete

3.51 MB

File too large to display, you can check the raw version instead.