Skip to content

[PyT] Reduce test sizes in fused attn fp8 vs fp16 to avoid OOM #3020

Open
vedaanta wants to merge 3 commits into
NVIDIA:mainfrom
vedaanta:vedaanta/te-fp8-vs-f16-shrink-b1
Open

[PyT] Reduce test sizes in fused attn fp8 vs fp16 to avoid OOM #3020
vedaanta wants to merge 3 commits into
NVIDIA:mainfrom
vedaanta:vedaanta/te-fp8-vs-f16-shrink-b1

Commits

Commits on May 27, 2026