Commit Graph

1022 Commits

Author SHA1 Message Date
liyonghua0910
e4a6710e4f Merge remote-tracking branch 'upstream/develop' into develop+clear_prefix_cache 2025-09-17 21:35:04 +08:00
gaoziyuan
896e3bb606 [NewFeture]add ep rollout model init and update/clear ep buffer (#4039)
* fix gid

* merge

* fix test

* fix bug

* fix

* fix ci
2025-09-17 20:24:53 +08:00
qw86972190
b52971749c Print KV Cache available memory and block memory usage in GB format (#4148) 2025-09-17 20:01:55 +08:00
liyonghua0910
c9cfd09939 [fix] fix code style 2025-09-17 19:23:19 +08:00
liyonghua0910
38d09a82c0 Merge remote-tracking branch 'upstream/develop' into develop+clear_prefix_cache 2025-09-17 19:10:29 +08:00
liyonghua0910
f62c4feeb8 [chore] add preemption triggered info log 2025-09-17 18:55:59 +08:00
liyonghua0910
33f209086b [fix] fix clear/update lock not working when workers > 1 2025-09-17 18:53:58 +08:00
liyonghua0910
cfa0982aae [fix] fix ep group all-reduce 2025-09-17 18:05:34 +08:00
RichardWooSJTU
2adca04f1f Reconstruct streaming data transfer with zmq (#3836)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
* reconstruct USE_GET_SAVE_OUTPUT_V1

* fix ut

* use dp rank

* fix ci
2025-09-17 14:30:39 +08:00
Jiang-Jia-Jun
f9766f917b [BugFix] Forbiden FD_DISABLED_RECOVER while ENABLE_V1_KVCACHE_SCHEDULER (#4142)
Co-authored-by: Jiang-Jia-Jun <jiangjiajun@baidu.com>
2025-09-17 14:11:44 +08:00
liyonghua0910
c33e362932 [fix] fix key/value_cache_scales indent 2025-09-17 11:42:51 +08:00
YuanRisheng
2e9e53ff7e [FDConfig]Remove max_num_batched_tokens/max_num_seqs in parallel config (#4116)
* remove max_num_batched_tokens in parallel config

* remove max_num_seqs

* update test case

* fix test

* fix

---------

Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
2025-09-17 10:43:35 +08:00
chenjian
67e6d8c691 [Feature] Set prefix caching as default (#3814)
* Set prefix caching as default

* Set prefix caching as default

* Set prefix caching as default

* skip dynamic load scene

* fix kill bug

* fix kill bug

* fix kill bug

* fix

* fix

* fix ci
2025-09-16 20:34:27 +08:00
Jiang-Jia-Jun
a04365a0c7 Update api_server.py 2025-09-15 21:31:33 +08:00
YuanRisheng
03b3d6175d fix mtp (#4105) 2025-09-15 20:26:07 +08:00
co63oc
17a27170bc fix typos (#4093) 2025-09-15 18:33:30 +08:00
bukejiyu
113e330030 fix bf16 and add comments (#4106) 2025-09-15 17:23:07 +08:00
freeliuzc
69aa2781a1 [MTP]Support mtp reshard (#4099)
* support rl reshard

* modify model name
2025-09-15 17:13:53 +08:00
freeliuzc
46911f903d [MTP]update hybrid-mtp-with-ngram (#4047) 2025-09-15 17:13:31 +08:00
Yuanle Liu
b1b33211e8 [CUDAGraph] Support multi output buffers and merge some fixes from feature/exp_0908 (#4062)
* refine cudagraph

* refine cudagraph

* typo

* fix

* fix plugins

* fix

* update

* update

* update
2025-09-15 16:21:30 +08:00
zhupengyang
9409665713 [xpu] support ep (#4067)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
2025-09-15 13:53:11 +08:00
bukejiyu
29ed617f0f [v1 loader]qwen Offline fp8 (#4036)
* support offline fp8

* update ut

* update ut

* update ut

* fix

* update

* update
2025-09-15 13:44:11 +08:00
Sunny-bot1
b1a5b756a3 [Optimize] Support WINT8 and group scale for Machete (#3905) 2025-09-15 12:01:34 +08:00
Zero Rains
f213ae1e86 [Bug Fix]fix the bug for cache_messager signal loss (#3879)
* fix the bug for real size 0 in cudagraph

* fix cache_messager
2025-09-15 11:16:24 +08:00
李泳桦
1411415816 Merge branch 'develop' into develop+clear_prefix_cache 2025-09-15 10:54:24 +08:00
qwes5s5
553adb299e 【FastDeploy CLI】collect-env subcommand (#4044)
* collect-env subcommand

* trigger ci

---------

Co-authored-by: K11OntheBoat <your_email@example.com>
2025-09-15 10:31:23 +08:00
zhouchong
958abebeab Support offline inference with streaming output (#4071)
* Support offline inference with streaming output

* add unit test

---------

Co-authored-by: Jiang-Jia-Jun <163579578+Jiang-Jia-Jun@users.noreply.github.com>
2025-09-15 10:27:03 +08:00
liyonghua0910
94a55fc158 [fix] fix prefix caching not enabled 2025-09-12 21:43:37 +08:00
liyonghua0910
c7b8f4f8c6 [fix] fix ipc suffix, use port instead 2025-09-12 21:41:59 +08:00
liyonghua0910
013338f2f5 [feat] support clearing prefix cache (cherry-picked from release/2.1) 2025-09-12 21:41:55 +08:00
Ayakouji
987609c894 [BugFix] Fix image_feature 0-Size causing insert failed (#4042)
* update

* fix image_feature
2025-09-12 19:13:08 +08:00
xiaolei373
9ac539471d [format] Valid para format error info (#4035)
* feat(log):add_request_and_response_log

* 报错信息与OpenAI对齐
2025-09-12 19:05:17 +08:00
YuanRisheng
88ea565aba [BugFix]Fix load kv cache quant scale (#4077)
* fix kv cache

* fix kv_cache

* fix kv cache
2025-09-12 17:44:03 +08:00
SuperNova
805f29a06c [Feature] refactor metax_gpu attention and moe and remove some useless code (#3688)
Co-authored-by: yongqiangma <xing.wo@163.com>
2025-09-12 14:40:25 +08:00
qwes5s5
58e0785bab [metrics] update metrics markdown file (#4061)
* adjust md

* trigger ci

---------

Co-authored-by: K11OntheBoat <your_email@example.com>
2025-09-12 11:13:43 +08:00
co63oc
8466219ec8 fix typos (#3840)
* fix typos

* ci

---------

Co-authored-by: YuBaoku <49938469+EmmonsCurse@users.noreply.github.com>
2025-09-12 11:04:38 +08:00
RichardWooSJTU
82dab8a91a Add token processor plugin support (#4059)
* Add token processor plugin support

* fix import

* fix import
2025-09-12 10:17:23 +08:00
chenjian
37f1632732 [Optimize] optimize prefix cache in develop (#3890)
* optimize prefix cache in release22

* fix

* fix

* fix

* add ci for v1

* add unit test

---------

Co-authored-by: xiegegege <46314656+xiegegege@users.noreply.github.com>
2025-09-12 10:15:59 +08:00
chen
4859f40b20 [Feature] GLM-45-AIR Support Mix Quantization(Dense wfp8afp8 and wint8 triton_moe_backend) (#4051)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
Publish Job / publish_pre_check (push) Has been cancelled
Publish Job / print_publish_pre_check_outputs (push) Has been cancelled
Publish Job / FD-Clone-Linux (push) Has been cancelled
Publish Job / Show Code Archive Output (push) Has been cancelled
Publish Job / BUILD_SM8090 (push) Has been cancelled
Publish Job / BUILD_SM8689 (push) Has been cancelled
Publish Job / PADDLE_PYPI_UPLOAD_8090 (push) Has been cancelled
Publish Job / PADDLE_PYPI_UPLOAD_8689 (push) Has been cancelled
Publish Job / Run FastDeploy Unit Tests and Coverage (push) Has been cancelled
Publish Job / Run FastDeploy LogProb Tests (push) Has been cancelled
Publish Job / Extracted partial CE model tasks to run in CI. (push) Has been cancelled
Publish Job / Run Base Tests (push) Has been cancelled
Publish Job / Run Accuracy Tests (push) Has been cancelled
Publish Job / Run Stable Tests (push) Has been cancelled
CI Images Build / FD-Clone-Linux (push) Has been cancelled
CI Images Build / Show Code Archive Output (push) Has been cancelled
CI Images Build / CI Images Build (push) Has been cancelled
CI Images Build / BUILD_SM8090 (push) Has been cancelled
CI Images Build / Run FastDeploy Unit Tests and Coverage (push) Has been cancelled
CI Images Build / Run FastDeploy LogProb Tests (push) Has been cancelled
CI Images Build / Extracted partial CE model tasks to run in CI. (push) Has been cancelled
CI Images Build / Run Base Tests (push) Has been cancelled
CI Images Build / Run Accuracy Tests (push) Has been cancelled
CI Images Build / Run Stable Tests (push) Has been cancelled
CI Images Build / Publish Docker Images Pre Check (push) Has been cancelled
2025-09-11 20:08:09 +08:00
lddfym
2056a428bd [bug fix] Fix the placeholder in qwen prompt and add some unittests (#4065)
* fix the placeholder in qwen prompt

* fix the placeholder in qwen prompt

* add soem unittests for qwen_vl_processor
2025-09-11 20:00:02 +08:00
memoryCoderC
850465e8ed [Feature] add cli command chat,complete (#4037) 2025-09-11 19:53:14 +08:00
zhuzixuan
a47976e82d [Echo] Support more types of prompt echo (#4022)
* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

* wenxin-tools-700 When the prompt type is list[int] or list[list[int]], it needs to support echoing after decoding.

---------

Co-authored-by: luukunn <83932082+luukunn@users.noreply.github.com>
2025-09-11 19:34:44 +08:00
xiaoxiaohehe001
abdcef30aa [BugFix] mm_post_fix (#4005)
* mm_post_fix

* mm_post_fix_1
2025-09-11 19:09:46 +08:00
CSWYF3634076
e4c64a71cc [BugFix] qwen2.5vl enable_thinking=true and image_patch_id bug fix (#3921) 2025-09-11 15:08:24 +08:00
bukejiyu
2650f58740 [docs] Update environment variables documentation (#3957)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
2025-09-10 21:17:06 -07:00
AIbin
a7392a0ff9 【Inference Optimize】DeepSeek-V3-model MLA Optimize (#3886)
* support MLA chunk_size auto search & cuda_graph
2025-09-11 10:46:09 +08:00
chen
637d96c6ae [Feature] Support zai-org/GLM-4.5-Air BF16 model (#3928)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
Publish Job / publish_pre_check (push) Has been cancelled
Publish Job / print_publish_pre_check_outputs (push) Has been cancelled
Publish Job / FD-Clone-Linux (push) Has been cancelled
Publish Job / Show Code Archive Output (push) Has been cancelled
Publish Job / BUILD_SM8090 (push) Has been cancelled
Publish Job / BUILD_SM8689 (push) Has been cancelled
Publish Job / PADDLE_PYPI_UPLOAD_8090 (push) Has been cancelled
Publish Job / PADDLE_PYPI_UPLOAD_8689 (push) Has been cancelled
Publish Job / Run FastDeploy Unit Tests and Coverage (push) Has been cancelled
Publish Job / Run FastDeploy LogProb Tests (push) Has been cancelled
Publish Job / Extracted partial CE model tasks to run in CI. (push) Has been cancelled
Publish Job / Run Base Tests (push) Has been cancelled
Publish Job / Run Accuracy Tests (push) Has been cancelled
Publish Job / Run Stable Tests (push) Has been cancelled
CI Images Build / FD-Clone-Linux (push) Has been cancelled
CI Images Build / Show Code Archive Output (push) Has been cancelled
CI Images Build / CI Images Build (push) Has been cancelled
CI Images Build / BUILD_SM8090 (push) Has been cancelled
CI Images Build / Run FastDeploy Unit Tests and Coverage (push) Has been cancelled
CI Images Build / Run FastDeploy LogProb Tests (push) Has been cancelled
CI Images Build / Extracted partial CE model tasks to run in CI. (push) Has been cancelled
CI Images Build / Run Base Tests (push) Has been cancelled
CI Images Build / Run Accuracy Tests (push) Has been cancelled
CI Images Build / Run Stable Tests (push) Has been cancelled
CI Images Build / Publish Docker Images Pre Check (push) Has been cancelled
* support glm45_air
2025-09-10 19:36:10 +08:00
ltd0924
684e93269b [Fix] fix multi api server log dir (#3967)
* [BugFix] fix max streaming tokens invalid

* fix scheduler bug

* fix scheduler bug

* Update multi_api_server.py
2025-09-10 17:15:30 +08:00
RAM
d3e4ae3d49 [Executor] Adjust signal sending order in RL training (#3773)
* Adjust processing order

* fix bug

* fix update_parameters bug

* refine code
2025-09-10 13:24:20 +08:00
Ayakouji
453487d5b0 [Feat] ernie4_5_vl_moe support CudaGraph (#3226)
* delete dynamic control flow for decode

* coda-style

* fix scatter/gather typos and use input stream instead default stream

* support 0-Size Tensor

* update runner and model

* using static mem address as input

* fix mem leak

* refine code

* update mm_buffer

* fix typo

* fix buffersize

* fix unk token

* refine code

* refine

* support other arch

* open cudagraph in vlci

* fix

* update

* update

* update

* fix cmd

* update

---------

Co-authored-by: aquagull <hongyuh@qq.com>
Co-authored-by: Yuanle Liu <yuanlehome@163.com>
2025-09-10 13:11:57 +08:00