4 people b3eac168e7 Support triton kernels v3.4.0 for fused_moe (#8258) ...

Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
Co-authored-by: Cheng Wan <cwan@x.ai>
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>

2025-07-27 02:28:49 -07:00

..

compressed_tensors

[code style] Clean dead triton kernel code in fused_moe and useless vllm_ops import (#8310)

2025-07-24 14:38:30 +08:00

configs

Add H20 tuning configs support DeepSeek V3/R1 INT8(block-wise) (#4220)

2025-03-11 12:32:33 -07:00

deep_gemm_wrapper

[fix] fix DeepGEMM blackwell input quant & ut & fix style and log (#7247)

2025-06-16 11:45:54 -07:00

__init__.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

awq_triton.py

[AMD] Add triton awq_dequantize kernel to support AWQ on ROCm (#7661)

2025-07-18 14:27:25 -07:00

awq.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

base_config.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

blockwise_int8.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

fp8_kernel.py

[AMD] Remove vllm's scaled_fp8_quant and moe_sum when SGLANG_USE_AITER=1 (#7484)

2025-07-21 17:33:19 -07:00

fp8_utils.py

[Dockerfile] Multi-arch support for ROCm (#7902)

2025-07-14 06:13:09 +00:00

fp8.py

[2/N] MoE Refactor: Unify weight loader and quant methods (#8397)

2025-07-27 01:00:21 -07:00

gptq.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

int8_kernel.py

[2/2] Support Qserve (#6521)

2025-05-23 12:39:18 -07:00

int8_utils.py

Feature DeepSeek V3/R1 INT8 Quantization (block-wise) (#3730)

2025-02-24 05:43:35 -08:00

kv_cache.py

[refactor] slightly tidy fp8 module (#5993)

2025-05-07 17:28:24 -07:00

marlin_utils.py

Refactor: move all quantization-related code to srt/layer/quantization (#7989)

2025-07-17 00:47:07 -07:00

modelopt_quant.py

Use FlashInfer FP4 gemm. (#8241)

2025-07-27 01:05:22 -07:00

moe_wna16.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

petit_utils.py

Support NVFP4 quantized dense models on AMD CDNA2/CDNA3 GPUs (#7302)

2025-07-18 19:59:39 -07:00

petit.py

[fix] fix modelopt fp4 on b200 (#8195)

2025-07-20 17:39:57 -07:00

qoq.py

Refactor: move all quantization-related code to srt/layer/quantization (#7989)

2025-07-17 00:47:07 -07:00

scalar_type.py

[1/n] chore: decouple quantization implementation from vLLM dependency (#7992)

2025-07-16 15:56:26 -07:00

unquant.py

Support triton kernels v3.4.0 for fused_moe (#8258)

2025-07-27 02:28:49 -07:00

utils.py

[code style] Clean dead triton kernel code in fused_moe and useless vllm_ops import (#8310)

2025-07-24 14:38:30 +08:00

w4afp8.py

[2/N] MoE Refactor: Unify weight loader and quant methods (#8397)

2025-07-27 01:00:21 -07:00

w8a8_fp8.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00

w8a8_int8.py

[1/N] MoE Refactor: refactor select_experts (#7966)

2025-07-19 00:51:15 -07:00