ViperEkura 0654b4b916 refactor: template combine kernel, fix mask bug, unify dispatch ...

- Template combine kernel, share macros, extract entry_utils helpers
- Fix mask indexing (pass stride not pre-multiplied base)
- Remove !p.use_mask — MMA handles mask

2026-07-15 21:44:17 +08:00

..

kernels

refactor: template combine kernel, fix mask bug, unify dispatch

2026-07-15 21:44:17 +08:00

tests

fix: uninitialized strides in decode test and wrong stride helper in paged test

2026-07-14 23:58:30 +08:00

__init__.py

feat: add optional CUDA kernel system (csrc/) + fused GQA decode attention

2026-07-06 12:09:58 +08:00

build.py

chore: bump version to 1.3.9

2026-07-12 21:04:20 +08:00