Gahow Wang c0f0b67510 test: eps=2e-3 for flash dQ/dK finite-diff (cuts f32 rounding term) ...

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

2026-06-17 23:17:44 +08:00

..

src

autodiff+model: flash-attention op + --flash opt-in wiring

2026-06-17 23:10:32 +08:00

tests

test: eps=2e-3 for flash dQ/dK finite-diff (cuts f32 rounding term)

2026-06-17 23:17:44 +08:00

build.rs

autograd: tape engine + grad accumulation

2026-06-15 15:44:17 +08:00

Cargo.toml

ops: differentiable autograd nodes + per-op grad-check tests

2026-06-15 15:53:55 +08:00