SijiaYang cb9d91ea8a feat: support DeepSeek-R1-W4AFP8 model with ep-moe mode (#7762) ...

Signed-off-by: yangsijia.614 <yangsijia.614@bytedance.com>

2025-07-07 14:47:21 -07:00

..

compressed_tensors

Fix: sync prepare_fp8_layer_for_marlin with latest vllm changes (#7648)

2025-06-30 21:51:01 -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] fix DeepGEMM blackwell input quant & ut & fix style and log (#7247)

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

__init__.py

feat: support DeepSeek-R1-W4AFP8 model with ep-moe mode (#7762)

2025-07-07 14:47:21 -07:00

awq.py

fix awq_dequantize import (#5669)

2025-04-23 01:36:26 -07:00

base_config.py

Support compressed tensors fp8w8a8 (#4743)

2025-03-26 13:21:25 -07:00

blockwise_int8.py

Fuse routed scaling factor in deepseek (#6970)

2025-06-08 15:24:24 -07:00

fp8_kernel.py

Move multimodal processors into a separate folder (#7581)

2025-06-27 11:58:24 -07:00

fp8_utils.py

Move multimodal processors into a separate folder (#7581)

2025-06-27 11:58:24 -07:00

fp8.py

feat: support DeepSeek-R1-W4AFP8 model with ep-moe mode (#7762)

2025-07-07 14:47:21 -07:00

gptq.py

Fix GPTQMarlinMoE (#7697)

2025-07-01 22:34:43 -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

modelopt_quant.py

FlashInfer NVFP4 MoE with EP & 2-stream shared expert (#7327)

2025-06-22 13:38:47 -07:00

moe_wna16.py

[CPU] add optimizations for INT8 and FP8 DeepSeek (#6769)

2025-06-27 19:04:29 -07:00

qoq.py

[2/2] Support Qserve (#6521)

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

quant_utils.py

[1/n] apply wna16marlin kernel in moe weight only quantization (#7683)

2025-07-01 23:21:25 -07:00

utils.py

Update python API of activation, topk, norm and rope and remove vllm dependency (#6614)

2025-06-17 22:11:50 -07:00

w4afp8.py

feat: support DeepSeek-R1-W4AFP8 model with ep-moe mode (#7762)

2025-07-07 14:47:21 -07:00

w8a8_fp8.py

Fuse routed scaling factor in deepseek (#6970)

2025-06-08 15:24:24 -07:00

w8a8_int8.py

[CPU] refine CPU integration code (#7647)

2025-07-03 09:51:09 -07:00