yangjianfengo1 93fcf7e4ec 【New Feature】W4afp8 supports per group quantization (#4272) ...

* w4afp8 支持per group

* code style

* 精度完成

* revert append attn utils

* ffn1 动态量化

* ffn2 支持动态量化

* code style

* code style

* 修改单测

* 修改单测

* fix bug

* Implement conditional parameter creation for layers

Add parameter creation for up_gate_proj_in_scale when ep_size > 1.

* code style

* fix conflict

* code style

* code style

* 修复w4aint8 精度

* fix ci

---------

Co-authored-by: yuanxiaolan <yuanxiaolan01@baidu.com>

2025-11-05 21:00:23 +08:00

..

fast_hardmard

【New Feature】W4afp8 supports per group quantization (#4272)

2025-11-05 21:00:23 +08:00

moe_wna16_marlin_utils

【Inference Optimize】Support automatic generation of marlin kernel (#3149)

2025-08-01 22:43:18 +08:00

deepgemm_preprocess.cu

polish code with new pre-commit rule (#2923)

2025-07-19 23:19:27 +08:00

ep_moe_expert_dispatch.cu

[NewFeture]add ep rollout model init and update/clear ep buffer (#4039)

2025-09-17 20:24:53 +08:00

fused_moe_helper.h

【New Feature】W4afp8 supports per group quantization (#4272)

2025-11-05 21:00:23 +08:00

fused_moe_imp_op.h

polish code with new pre-commit rule (#2923)

2025-07-19 23:19:27 +08:00

fused_moe_op.h

【New Feature】W4afp8 supports per group quantization (#4272)

2025-11-05 21:00:23 +08:00

fused_moe.cu

refactor rl get_name_mappings_to_training (#2847)

2025-07-15 07:31:42 -07:00

gptq_marlin_repack.cu

Sync v2.0 version of code to github repo

2025-06-29 23:29:37 +00:00

group_swiglu_with_masked.cu

[LLM] First commit the llm deployment code

2025-06-09 19:20:15 +08:00

group_swiglu_with_masked.h

[LLM] First commit the llm deployment code

2025-06-09 19:20:15 +08:00

moe_deepgemm_depermute.cu

[LLM] First commit the llm deployment code

2025-06-09 19:20:15 +08:00

moe_deepgemm_permute.cu

polish code with new pre-commit rule (#2923)

2025-07-19 23:19:27 +08:00

moe_dispatch.cu

【New Feature】W4afp8 supports per group quantization (#4272)

2025-11-05 21:00:23 +08:00

moe_expert_ffn_wint2.cu

【New Feature】W4afp8 supports per group quantization (#4272)

2025-11-05 21:00:23 +08:00

moe_fast_hardamard_impl_common.h

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

moe_fast_hardamard_impl.cuh

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

moe_fast_hardamard_kernel.cu

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

moe_fast_hardamard_kernel.h

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

moe_ffn.cu

【New Feature】W4afp8 supports per group quantization (#4272)

2025-11-05 21:00:23 +08:00

moe_reduce.cu

[Feat] ernie4_5_vl_moe support CudaGraph (#3226)

2025-09-10 13:11:57 +08:00

moe_redundant_topk_select.cu

topk_gating_softmax support bias (#3405)

2025-08-15 11:57:45 +08:00

moe_topk_select.cu

topk_gating_softmax support bias (#3405)

2025-08-15 11:57:45 +08:00

moe_wna16_marlin_gemm.cu

Sync v2.0 version of code to github repo

2025-06-29 23:29:37 +00:00

moe_wna16_marlin_gemm.h

Sync v2.0 version of code to github repo

2025-06-29 23:29:37 +00:00

swigluoai.cu

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

swigluoai.h

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

template_config.json

【Hackathon 9th No.86】autogen MoeFastHardamardImplWrapper template_instantiation (#4592)

2025-10-30 10:28:36 +08:00

tritonmoe_preprocess.cu

Support GPT-OSS-BF16 (#4240)

2025-10-20 14:44:58 +08:00

winx_unzip.cu

rename fused_get_rope.cu (#3752)

2025-09-03 10:54:34 +08:00