Jesse Gross 05e0f21bec mlx: fuse sigmoid router head in glm4_moe_lite ...

DeepSeek-V2-style aux-loss-free routing computes sigmoid(gates) once but
needs it twice: the raw sigmoid output is gathered after top-k, while the
post-bias negation is the argpartition key. Fuse into a single multi-output
Compiled kernel returning both, saving two launches on the routing path
per token. Exposed as a general SigmoidRouter since the same pattern is
shared across DeepSeek-V2 descendants.

Improves glm4.7 generation performance by approximately 1%.

2026-04-20 15:02:14 -07:00

..

glm4_moe_lite.go

mlx: fuse sigmoid router head in glm4_moe_lite

2026-04-20 15:02:14 -07:00

parser_test.go

MLX: add header vendoring and remove go build tag (#14642)

2026-03-09 17:24:45 -07:00

parser.go

MLX: add header vendoring and remove go build tag (#14642)

2026-03-09 17:24:45 -07:00

render_test.go

MLX: add header vendoring and remove go build tag (#14642)

2026-03-09 17:24:45 -07:00

render.go

MLX: add header vendoring and remove go build tag (#14642)

2026-03-09 17:24:45 -07:00