3 people 982db4ebac Feat: GLM-4.6 supports shared experts fusion (#13873) ...

Signed-off-by: UranusSeven <109661872+UranusSeven@users.noreply.github.com>
Co-authored-by: Kevin-XiongC <kevin_xiong1997@outlook.com>
Co-authored-by: Mingyi Jin <jinmingyi1998@sina.cn>

2025-12-01 11:33:18 +08:00

..

configs

Feat: GLM-4.6 supports shared experts fusion (#13873)

2025-12-01 11:33:18 +08:00

__init__.py

[code style] restruct fused_moe to avoid very long single file (#9878)

2025-09-02 11:04:27 +08:00

fused_marlin_moe.py

[kimi k2 thinking] Avoid useless torch.zeros_ (#13596)

2025-11-21 13:15:27 +08:00

fused_moe_triton_config.py

Feat: GLM-4.6 supports shared experts fusion (#13873)

2025-12-01 11:33:18 +08:00

fused_moe_triton_kernels.py

[AMD] Enable fused shared expert append and flatten quant for fp8 deepseekR1 model (#13705)

2025-11-21 02:48:28 -08:00

fused_moe.py

Feat: GLM-4.6 supports shared experts fusion (#13873)

2025-12-01 11:33:18 +08:00

layer.py

fix RuntimeError: RMSNorm failed with error code an illegal memory access was encountered (#14135)

2025-11-29 12:17:41 -08:00

moe_align_block_size.py

[opt kimi k2 4 / n] Delete useless pad kernel in sgl_moe_align_block_size (#13587)

2025-11-21 13:16:42 +08:00

triton_kernels_moe.py

Refactor Triton-kernel MoE runner integration (#11795)

2025-10-23 18:47:28 -07:00