Skip to content
Back to pull request #53

perf: turbo VEC flash attention — +9% decode on CUDA via autoresearch #60

Sign in to view logs

perf: turbo VEC flash attention — +9% decode on CUDA via autoresearch

perf: turbo VEC flash attention — +9% decode on CUDA via autoresearch #60

Annotations

1 error

The logs for this run have expired and are no longer available.