Skip to content

[None][feat] Kimi k3 fuse decode attention residual tail - #17194

Draft
xguannv wants to merge 5 commits into
NVIDIA:mainfrom
xguannv:xguan/kimi_k3_attn_res_decode_fusion
Draft

xguannv wants to merge 5 commits into
NVIDIA:mainfrom
xguannv:xguan/kimi_k3_attn_res_decode_fusion

[TRTLLM][Kimi K3] Drop the one-off attn-res add+rmsnorm microbenchmark

3cea263
Select commit
Loading
Failed to load commit list.