kk d4bf5a8524 Support OCP MXFP4 quantization on AMD GPUs (#8255) ...

Co-authored-by: wunhuang <wunhuang@amd.com>
Co-authored-by: Hubert Lu <Hubert.Lu@amd.com>

2025-08-04 18:14:52 -07:00

..

compressed_tensors

[hotfix] fix mixtral with tensor-level compressed-tensor quantization (#8721)

2025-08-02 22:59:25 -07:00

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 deepgemm masked grouped gemm jit compile (#8679)

2025-08-01 15:37:59 -07:00

quark

Support OCP MXFP4 quantization on AMD GPUs (#8255)

2025-08-04 18:14:52 -07:00

__init__.py

Support OCP MXFP4 quantization on AMD GPUs (#8255)

2025-08-04 18:14:52 -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

fp4.py

Support OCP MXFP4 quantization on AMD GPUs (#8255)

2025-08-04 18:14:52 -07:00

fp8_kernel.py

Remove assertions about per group quant fp8 (#8717)

2025-08-02 17:08:40 -07:00

fp8_utils.py

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

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

fp8.py

feat: support cutlass_moe_fp8 kernel for fusedmoe in sm90 (#8678)

2025-08-03 10:47:15 -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

[bugfix] Fix typo in modelopt quant: 'FusedMoE' object has no attribute 'local_num_experts' (#8768)

2025-08-04 11:08:08 -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

[4/N] MoE Refactor: Unified Triton Kernel for FusedMoE and EPMoE (#8515)

2025-07-31 02:34:02 -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

Fix bug of refactoring TopKOutput in w4afp8 (#8745)

2025-08-03 20:05:02 -07:00

w8a8_fp8.py

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

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

w8a8_int8.py

[Bugfix] fix w8a8_int8 load issue (#8308)

2025-07-31 17:30:16 -07:00