Skip to content

fix(deepseek-v4): correct V4-Flash attention and MoE math, and make the 4N recipe reproducible - #930

Open
lhzhang333 wants to merge 18 commits into
dev/tas/deepseek-v4-phase2from
dev/lhz/dsv4_fix
Open

fix(deepseek-v4): correct V4-Flash attention and MoE math, and make the 4N recipe reproducible#930
lhzhang333 wants to merge 18 commits into
dev/tas/deepseek-v4-phase2from
dev/lhz/dsv4_fix

refactor(deepseek-v4): drop the memory-reporting knobs that never rea…

0645cc4
Select commit
Loading
Failed to load commit list.

Select a check to view from the sidebar