Build uploaded using `kernels`.

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Revert "Build uploaded using `kernels`."

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Build uploaded using `kernels`.

Update tag

Build (aarch64-linux)

Build (x86_64-linux)

Update flake

Build (x86_64-linux Torch 2.8)

Build (aarch64-linux Torch 2.8)

Build (x86_64-linux)

Build (aarch64-linux)

Enable ROCm build

Sync CUDA paged-attention with upstream

Improve tests for mps

Update readme

Add metal paged attention

feat: add tag for hfjob build

Build

Build (AArch64)

Update flake inputs

Use default CUDA capabilities

Fix flake input

Build (aarch64)

Build

Sync capabilities with upstream

Update flake

feat: update to include rev in kernel for reproducible symbols

Update for build.toml changes

Update README.md (#1)

Build

Rename to paged-attention

Build

Port vLLM attention kernels

initial commit