zhupengyang 8735cb5045 [XPU] refactor moe ffn (#5501) ...

- remove BKCL_DISPATCH_ALL_GATHER
- support sparse mode
- support moe quant_method

2025-12-18 14:14:05 +08:00

..

__init__.py

support w4afp8 EP inference (#3044)

2025-08-25 11:27:45 +08:00

ep.py

[Others] update tbo related code (#5485)

2025-12-11 12:34:46 +08:00

fused_moe_backend_base.py

RL fix (#5503)

2025-12-11 19:25:27 +08:00

fused_moe_cutlass_backend.py

support w4afp8 mtp (#5429)

2025-12-08 20:24:00 +08:00

fused_moe_deepgemm_backend.py

[New][RL] Support Rollout Routing Replay (#5405)

2025-12-05 22:06:26 +08:00

fused_moe_marlin_backend.py

[New][RL] Support Rollout Routing Replay (#5405)

2025-12-05 22:06:26 +08:00

fused_moe_triton_backend.py

[New][RL] Support Rollout Routing Replay (#5405)

2025-12-05 22:06:26 +08:00

fused_moe_wint2_backend.py

[New][RL] Support Rollout Routing Replay (#5405)

2025-12-05 22:06:26 +08:00

moe.py

[XPU] refactor moe ffn (#5501)

2025-12-18 14:14:05 +08:00

routing_indices_cache.py

[RL] R3 Support RDMA Store (#5467)

2025-12-16 16:50:13 +08:00

triton_moe_kernels.py

[OPs] MoE support wfp8afp8(channelwise) and improve per_token_quant_fp8 (#4238)

2025-09-24 16:39:51 +08:00