Jesse Gross e1e3cec8d0 models: fuse MLP activation functions via mlx_compile ...

Converts SiLU/GELUApprox to compiled kernels and adds SwiGLU,
matching upstream mlx/mlx_lm's activations pattern. Routes llama,
qwen3, qwen3_5 (dense + MoE), and glm4_moe_lite MLP paths through
mlx.SwiGLU so each MLP invocation runs as one fused Metal/CUDA
kernel rather than a chain of per-op launches.

2026-04-14 16:38:32 -07:00

..

qwen3_5_test.go

mlx: add mxfp4/mxfp8/nvfp4 importing (#15015)

2026-03-24 13:45:44 -07:00

qwen3_5.go

models: fuse MLP activation functions via mlx_compile

2026-04-14 16:38:32 -07:00