Commit Graph

  • f50988d917 [Cherry-Pick][CI] Revert adapt vl_model baseline changes due to Paddle update(#5732) (#5733) release/2.4 YuBaoku 2025-12-24 12:14:34 +08:00
  • 06b772df22 Deployed 672620c with MkDocs version: 1.6.1 gh-pages 2025-12-24 03:59:51 +00:00
  • 672620cdfe Revert "[CI] Adapt vl_model baseline changes due to Paddle update (#5576)" (#5732) develop YuBaoku 2025-12-24 11:59:27 +08:00
  • 922a73ddd6 [Others] clean code (#5691) 周周周 2025-12-24 11:28:47 +08:00
  • f4f7224bed Revert "[CI] Adapt vl_model baseline changes due to Paddle update (#5576)" revert-5576-fix_unit_5 YuBaoku 2025-12-24 11:23:53 +08:00
  • 23d488c488 [Feature] Entropy calculation support (#5692) GoldPancake 2025-12-23 21:19:47 +08:00
  • d1c6e57341 [Others] upgrade paddleformer to 0.4.0 (#5599) bukejiyu 2025-12-23 21:08:01 +08:00
  • 85db9d5e56 [Others] reschedule preempt task support optional func (#5649) ming1753 2025-12-23 20:45:52 +08:00
  • 5badb9df39 Use consistent signal name resolution across all files copilot/optimize-process-exit-mechanism copilot-swe-agent[bot] 2025-12-23 12:16:43 +00:00
  • 9e4eb339b8 Address code review feedback copilot-swe-agent[bot] 2025-12-23 12:15:24 +00:00
  • 92d8b84236 修复剩余的"is not support"语法错误 copilot/fix-error-logs-syntax copilot-swe-agent[bot] 2025-12-23 12:14:19 +00:00
  • cd62cc2df9 Improve error handling in cleanup functions copilot-swe-agent[bot] 2025-12-23 12:13:11 +00:00
  • b68b7c8688 Add signal handling test guide documentation copilot-swe-agent[bot] 2025-12-23 12:11:42 +00:00
  • e09d6363d3 Add signal handlers for graceful process termination copilot-swe-agent[bot] 2025-12-23 12:10:20 +00:00
  • a79dfc108c 修复更多错误日志中的语法和拼写问题 copilot-swe-agent[bot] 2025-12-23 12:10:17 +00:00
  • f9a1233f11 Initial plan copilot-swe-agent[bot] 2025-12-23 12:03:52 +00:00
  • cee3a7a356 修复错误日志中缺少空格的问题 copilot-swe-agent[bot] 2025-12-23 11:59:18 +00:00
  • 73749c7bf3 修复错误日志中的非标准语法:"got error"和"meets error" copilot-swe-agent[bot] 2025-12-23 11:57:55 +00:00
  • f66d013f8d 修复代码中的日志错误消息:拼写和语法问题 copilot-swe-agent[bot] 2025-12-23 11:56:33 +00:00
  • c81211e219 Initial plan copilot-swe-agent[bot] 2025-12-23 11:52:26 +00:00
  • 5cec66adb8 [Docs] 更新环境变量文档以同步最新代码 (#5713) Copilot 2025-12-23 19:49:20 +08:00
  • 99258e19c8 [Benchmark]支持Completions接口 (#5700) ophilia-lee 2025-12-23 19:46:23 +08:00
  • 04c30521dd [Others] plugin raise error msg (#5675) ming1753 2025-12-23 18:56:54 +08:00
  • f15edbb6ef [CI]【Hackathon 9th Sprint No.40】功能模块 fastdeploy/entrypoints/openai/api_server.py 单测补充 (#5567) kesmeey 2025-12-23 18:06:43 +08:00
  • 9ff99d2b03 [BugFix] fix double shutdown of comm group when rank0 clears weights slower than other ranks (#5710) Yonghua Li 2025-12-23 17:51:35 +08:00
  • e9f5397bc9 [Docs] Update parameters documentation with latest code defaults and new parameters (#5709) Copilot 2025-12-23 17:31:44 +08:00
  • 5a74ee77f1 save model gzy19990617-patch-1 gaoziyuan 2025-12-23 16:30:13 +08:00
  • 1b74540820 fix eplb weight updating (#5529) (#5661) release/online/20251131 RichardWooSJTU 2025-12-23 16:09:05 +08:00
  • cfddec7142 [Quantization][Cherry-Pick] Support w4afp8 moe weight offline permute & load and DeepEP low latency two stage(#5613 #5608) (#5677) Sunny-bot1 2025-12-23 16:04:08 +08:00
  • 6250c686cc Revert "Revert "[Optim] Remove limitation of number of kvcache blocks (#5612)…" revert-5702-revert-5612-feature/remove_kvblock_limit_20251217 Jiang-Jia-Jun 2025-12-23 15:42:10 +08:00
  • c1aa66df02 Revert "[Optim] Remove limitation of number of kvcache blocks (#5612)" (#5702) Divano 2025-12-23 15:41:33 +08:00
  • 0bef9b684f [Metax][CI]fix CI bug (#5698) Jiaxin Sui 2025-12-23 14:56:34 +08:00
  • 2c3c983b96 [XPU] modify speculate_verify (#5522) RuohengMa 2025-12-23 14:50:30 +08:00
  • 945a1bc4e2 [Metax] update ci name (#5679) MingkunZhang 2025-12-23 14:00:48 +08:00
  • 6c36a17369 [Others]Prevent core dumps during Paddle version check (#5657) bukejiyu 2025-12-23 13:57:45 +08:00
  • 52280bee61 [Speculative Decoding]Support multi-step mtp with cudagraph (#5624) (#5695) freeliuzc 2025-12-23 13:21:59 +08:00
  • ceafd757f0 [Speculative Decoding]Support multi-step mtp with cudagraph (#5624) (#5670) freeliuzc 2025-12-23 13:18:47 +08:00
  • 9da89a374b [Optim] Remove limitation of number of kvcache blocks (#5612) Jiang-Jia-Jun 2025-12-23 11:18:29 +08:00
  • 4a74f5ab9b [XPU]Set top_p=0.0 by default on XPU to optimize performance (#5686) ddchenhao66 2025-12-23 11:01:01 +08:00
  • eb309e5a2a [XPU]Set top_p=0.0 by default on XPU to optimize performance (#5688) ddchenhao66 2025-12-23 11:00:53 +08:00
  • 3aee5c4bf5 [CI] 【Hackathon 9th Sprint No.37】NO.37 功能模块单测补充 (#5059) xunyoyo 2025-12-23 10:35:16 +08:00
  • f16077a939 [XPU][CI] Xpu ci update (#5690) Jiaxin Sui 2025-12-23 10:19:39 +08:00
  • dfe8ea941c [log]console log to llm log (#5680) xiaolei373 2025-12-23 10:05:45 +08:00
  • 131defa122 Revert "Revert "[Feature] Use paddle.compat.enable_torch_proxy in `fastdepl…" (#5606) RAM 2025-12-22 22:37:51 +08:00
  • a1535c7e7e [XPU][CI] xpu add ci test for pd + TP2 (#5653) ddchenhao66 2025-12-22 19:27:10 +08:00
  • 8beb0158fa [BugFix] fix rl signal (#5681) Yuanle Liu 2025-12-22 16:35:54 +08:00
  • 90065084cb [BugFix] fix rl signal (#5678) Yuanle Liu 2025-12-22 16:31:24 +08:00
  • 6ed9136a4e [Metax] update ci yaml (#5674) MingkunZhang 2025-12-22 16:00:25 +08:00
  • b57deb671d [CI] Update check_approval.sh YuBaoku 2025-12-22 15:52:04 +08:00
  • 04035e4ebf support w4afp8 two stage (#5608) Sunny-bot1 2025-12-22 15:13:05 +08:00
  • 40f3897a4e support w4afp8 moe offline permute & load (#5613) Sunny-bot1 2025-12-22 15:12:57 +08:00
  • 81384ef29e [BugFix] fix download feature bug (#5669) ming1753 2025-12-22 13:46:39 +08:00
  • 6d323769dd fix w4afp8 (#5634) lizexu123 2025-12-22 13:39:41 +08:00
  • 6eada4929d [Speculative Decoding]Support multi-step mtp with cudagraph (#5624) freeliuzc 2025-12-22 11:34:04 +08:00
  • ea16c82b43 [Cherry-Pick] [RL] provide options for whether shutdown comm group after weights cleared (#5663) (#5664) Yonghua Li 2025-12-19 23:18:03 +08:00
  • 4f830aa505 [RL] provide options for whether shutdown comm group after weights cleared (#5663) Yonghua Li 2025-12-19 23:06:48 +08:00
  • fe55baae47 [CI] Fix unit_test error of unstable execution (#5660) YuBaoku 2025-12-19 22:59:53 +08:00
  • abf53b17ea [BugFix] Fix custom_all_reduce overflow (#5662) (#5667) chen 2025-12-19 20:04:39 +08:00
  • a32cb54d0b [BugFix] Fix custom_all_reduce overflow (#5662) chen 2025-12-19 18:24:21 +08:00
  • 46d83be065 [Metax] update ci test (#5652) MingkunZhang 2025-12-19 17:25:47 +08:00
  • dd0014b7b9 del core (#5659) bukejiyu 2025-12-19 16:33:44 +08:00
  • 669dfe8dca [CI] 【Hackathon 9th Sprint No.38】NO.38 功能模块单测补充 (#5060) xunyoyo 2025-12-19 16:28:16 +08:00
  • 807e404369 [BugFix] fix eb5 mm prefix cache bug (#5638) kevin 2025-12-19 14:57:37 +08:00
  • e10c5d5d61 cp fix eb5 prefix cache bug (#5644) kevin 2025-12-19 14:57:17 +08:00
  • 6bd772b93f fix eplb weight updating (#5529) RichardWooSJTU 2025-12-19 14:30:32 +08:00
  • a9bb24bb56 [XPU]logprob bug (#5636) qw86972190 2025-12-19 14:30:14 +08:00
  • 689f54f671 [RL] Update worker_process.py (#5651) Yuanle Liu 2025-12-19 12:07:58 +08:00
  • b3f78815d8 update rl signal (#5650) Yuanle Liu 2025-12-19 12:04:18 +08:00
  • a8fce47195 [Intel HPU] enable kv cache scheduler v1 for hpu (#5648) fmiao2372 2025-12-19 12:03:39 +08:00
  • 23bfd28624 [Cherry-Pick][BugFix] cp fix_cpu_cache_bugs(#5544) (#5577) kevin 2025-12-19 11:48:50 +08:00
  • 2aa88d3621 [Cherry-Pick][RL]Fix RL load_weights #5642 (#5643) bukejiyu 2025-12-19 11:17:09 +08:00
  • fc452c8e29 [RL]Fix RL load_weights (#5642) bukejiyu 2025-12-19 11:16:18 +08:00
  • ec6811f648 support token num = 0 (#5635) lizan1999 2025-12-19 10:20:38 +08:00
  • d657455616 [CI] 【Hackathon 9th Sprint No.19】NO.19 功能模块单测补充 (#5063) xunyoyo 2025-12-18 21:32:44 +08:00
  • 9c55bc31cd [Cherry-Pick][BugFix] fix rl model_weights_signal to support tp>1 #5639 (#5637) Yuanle Liu 2025-12-18 20:44:19 +08:00
  • b47674c796 [BugFix] fix rl model_weights_signal to support tp>1 (#5639) Yuanle Liu 2025-12-18 20:43:58 +08:00
  • d739af5e6e Revert "[XPU][CI] xpu add ci test for pd (#5610)" (#5645) Jiaxin Sui 2025-12-18 19:59:09 +08:00
  • 646d1a0aa2 [Cherry-Pick][RL]Support loading weights via the load_weights function for RL #5549 (#5602) bukejiyu 2025-12-18 18:28:53 +08:00
  • 4aa2c6871b [RL]Support loading weights via the load_weights function for RL (#5549) bukejiyu 2025-12-18 18:27:05 +08:00
  • ac013803f3 [Iluvatar] Support V1_KVCACHE_SCHEDULER and paddleocr-vl rope mode (#5555) yzwu 2025-12-18 18:14:25 +08:00
  • 0cb9ad186e [Cherry-Pick][BugFix] fix speculate_limit_thinking_content_length #5590 (#5615) Yuanle Liu 2025-12-18 17:50:18 +08:00
  • 48f3e9797e Update backend_request_func.py (#5633) Zhang Yulong 2025-12-18 16:21:34 +08:00
  • 2d2619d300 [CI] 【Hackathon 9th Sprint No.36】NO.36 功能模块单测补充 (修复) (#5609) xunyoyo 2025-12-18 16:08:42 +08:00
  • a30a5b4216 [Model] tp+ep support v1_loader (#5600) Longzhi Wang 2025-12-18 15:27:12 +08:00
  • e1a9b282eb fix bug for EP+MTP (#5605) lizan1999 2025-12-18 14:34:54 +08:00
  • d8587e987e [Model] tp+ep support v1_loader (#5465) Longzhi Wang 2025-12-18 14:31:54 +08:00
  • c89a62e550 Update backend_request_func.py (#5631) Zhang Yulong 2025-12-18 14:20:17 +08:00
  • 8735cb5045 [XPU] refactor moe ffn (#5501) zhupengyang 2025-12-18 14:14:05 +08:00
  • d0a7834a17 [Metax] fix metax runner issue (#5629) MingkunZhang 2025-12-18 13:32:54 +08:00
  • c606df59f5 [XPU]logprob bug (#5626) qw86972190 2025-12-18 12:07:20 +08:00
  • d81341b9b3 [CI]【Hackathon 9th Sprint No.14】功能模块 fastdeploy/rl/rollout_model.py 单测补充 (#5552) kesmeey 2025-12-18 10:57:53 +08:00
  • 5300e73f8b [Others] Maintain the mtp branch temporarily. (#5446) (#5621) lzy 2025-12-17 22:03:25 +08:00
  • f45c131ddf update (#5625) Zhang Yulong 2025-12-17 21:38:14 +08:00
  • 94be5ebdd1 [CI] Add CI case for MTP accept ratio (#5570) Zhang Yulong 2025-12-17 21:35:02 +08:00
  • e56c4dd0a8 [Cherry-Pick] Support for request-level speculative decoding metrics monitoring.(#5518) (#5614) GoldPancake 2025-12-17 20:53:04 +08:00
  • ac731653b3 [CI]【Hackathon 9th Sprint No.12】功能模块 fastdeploy/spec_decode/mtp.py 单测补充 (#5533) kesmeey 2025-12-17 20:09:45 +08:00
  • d7d633a285 [Cherry-Pick][CI]Fix write qknorm cache bug in speculative decoding(#5491) (#5617) freeliuzc 2025-12-17 20:08:51 +08:00
  • 53f4a9ad27 Simplify implementation: use inline acquire/release with shared memory counter copilot/fix-concurrency-control-issue copilot-swe-agent[bot] 2025-12-17 11:09:17 +00:00
  • 111955ec0c [BugFix] 移除重复的 PaddleOCRVLProcessor 初始化代码 megemini 2025-12-12 13:26:41 +08:00
  • 6f9b25902a Address code review feedback: improve IPCSignal initialization, remove unused function, fix formatting copilot-swe-agent[bot] 2025-12-17 10:56:26 +00:00