yangjianfengo1 ae7bee8122 【New Feature】W4afp8 supports per group quantization (#4987) ...

* w4afp8 支持per group

* code style

* fix transpose

* revert fast hardmard

---------

Co-authored-by: yuanxiaolan <yuanxiaolan01@baidu.com>
Co-authored-by: plusNew001 <95567040+plusNew001@users.noreply.github.com>

2025-11-13 19:17:27 +08:00

..

benchmarks

[CLI]Update parameters in bench latecy cli tool and fix collect-env cli tool (#4558)

2025-10-24 16:46:45 +08:00

cache_manager

[KVCache] support unified cache backend (#4903)

2025-11-12 14:54:52 +08:00

demo

[KVCache] support unified cache backend (#4903)

2025-11-12 14:54:52 +08:00

distributed

fix import jit.marker.unified (#4622)

2025-10-28 22:11:03 +08:00

engine

[BugFix] fix num_requests_running after clear_data (#4927)

2025-11-13 13:50:21 +08:00

entrypoints

[Logprobs]Support prompt_logprobs and max_logprobs (#4897)

2025-11-12 19:29:48 +08:00

eplb

[Feature] Support eplb for fd (#4599)

2025-11-03 14:08:15 +08:00

input

[TSP] Support qwen3 moe tsp + cudagraph (#4871)

2025-11-10 23:37:51 +08:00

inter_communicator

[Perf] Support tensor transmission between work and engine with zero-copy to improve efficiency (#4839)

2025-11-11 15:43:11 +08:00

logger

[BugFix] PaddleOCR-VL fix FD_DEBUG type and support v1 loader (#4605)

2025-10-28 09:47:47 +08:00

metrics

[fix] fix requests & block metrics (#4404)

2025-10-15 11:49:24 +08:00

model_executor

【New Feature】W4afp8 supports per group quantization (#4987)

2025-11-13 19:17:27 +08:00

multimodal

[Bug Fix] process transparent image (#4807)

2025-11-05 17:15:24 +08:00

output

[BugFix] fix num_requests_running after clear_data (#4927)

2025-11-13 13:50:21 +08:00

platforms

[Metax] adapt DeepSeek (#4498)

2025-10-24 10:14:53 +08:00

plugins

[Feature] add cli command serve (#4226)

2025-09-24 14:50:45 +08:00

reasoning

[BugFix] Fix ernie_vl_reasoning_parsers.py 'end_token' to 'think_end_token' (#4805)

2025-11-06 11:28:55 +08:00

rl

[Speculative Decoding][MTP]Support mtp in epdptp mode (#4614)

2025-10-28 16:02:47 +08:00

router

[Feature] [PD] add simple router and refine splitwise deployment (#4709)

2025-11-06 14:56:02 +08:00

scheduler

[Feature] [PD] add simple router and refine splitwise deployment (#4709)

2025-11-06 14:56:02 +08:00

spec_decode

[KVCache] support unified cache backend (#4903)

2025-11-12 14:54:52 +08:00

splitwise

[Bug Fix] fix bug for PD EP (#4823)

2025-11-10 15:33:29 +08:00

transformer_utils

[Feature] support pool (#3827)

2025-09-22 14:09:09 +08:00

worker

[worker_process.py]modify some var name (#4749)

2025-11-13 14:21:27 +08:00

__init__.py

[BugFix] PaddleOCR-VL fix FD_DEBUG type and support v1 loader (#4605)

2025-10-28 09:47:47 +08:00

collect_env.py

feat: add support for API usage with multimodal models (#4548)

2025-10-28 20:23:46 +08:00

config.py

[FDConfig] add block number verfied (#4983)

2025-11-13 09:48:44 +08:00

envs.py

[TSP] Support qwen3 moe tsp + cudagraph (#4871)

2025-11-10 23:37:51 +08:00

import_ops.py

fix import error (#2944)

2025-07-22 14:06:01 +08:00

stop.sh

[PD Disaggregation] Support Qwen3-MoE use PD + EP inference. (#4691)

2025-11-06 10:32:15 +08:00

test.yaml

[Sync] Update to latest code (#2679)

2025-07-03 15:43:53 +08:00

utils.py

[Feature] [PD] add simple router and refine splitwise deployment (#4709)

2025-11-06 14:56:02 +08:00