lizexu123 44a13e4557 [Feature] support w4afp8 v1_loader and v0_loader(tp>1) (#5757) ...

* support

* fix

* support w4afp8 v1_loader and v0_loader

* fix

* fix test

* fix test

* fix test

* fix moe.py

* add test_ernie_4_5_w4afp8

* add test

* delete tensor

* fix test

* fix

* add

* fix test

2025-12-30 14:11:52 +08:00

..

__init__.py

support w4afp8 EP inference (#3044)

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

ep.py

support w4afp8 two stage (#5608)

2025-12-22 15:13:05 +08:00

fused_moe_backend_base.py

RL fix (#5503)

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

fused_moe_cutlass_backend.py

[Feature] support w4afp8 v1_loader and v0_loader(tp>1) (#5757)

2025-12-30 14:11:52 +08:00

fused_moe_deepgemm_backend.py

[BugFix] Fix return value inconsistency for ep_moe_expert_combine op (#5812)

2025-12-29 16:44:00 +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

[GraphOptimization] Wrap deep gemm and triton as python op (#5673)

2025-12-24 15:23:46 +08:00

fused_moe_wint2_backend.py

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

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

moe.py

support w4afp8 moe offline permute & load (#5613)

2025-12-22 15:12:57 +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