Skip to content

[CUDA] Support FP8 (E4M3) KV Cache for Group Query Attention #20665

[CUDA] Support FP8 (E4M3) KV Cache for Group Query Attention

[CUDA] Support FP8 (E4M3) KV Cache for Group Query Attention #20665

Annotations

2 warnings

webgpu  /  build-and-test (arm64, arm64, Release)

succeeded Feb 14, 2026 in 59m 20s