fraQtl
/

Mistral-7B-optimized

kv-cache-optimized

Model card Files Files and versions

Mistral-7B-optimized

Commit History

Add arXiv paper link

c367f9b
verified

Zenalyze commited on 1 day ago

Clarify: KV cache optimized, not smaller file

9a7a97c
verified

Zenalyze commited on 2 days ago

Update: k=64, fix generation samples, correct PPL

dd72361
verified

Zenalyze commited on 4 days ago

fraQtl compressed: k=64 INT3, delta=+0.2218

b881508
verified

Zenalyze commited on 4 days ago

fraQtl compressed: k=48 INT3, delta=+0.2146

6fe65fa
verified

Zenalyze commited on 5 days ago

Update model card with runtime comparison

80efb64
verified

Zenalyze commited on 5 days ago

fraQtl compressed: k=32 INT3, delta=+0.1989

dc721d2
verified

Zenalyze commited on 5 days ago

initial commit

befc852
verified

Zenalyze commited on 5 days ago