Commit Graph

88 Commits

Author SHA1 Message Date
Yuge Zhang 4e87832970 Merge branch 'main' of github.com:microsoft/agent-lightning into apo-algorithm 2025-10-14 09:03:39 +08:00
Yuge Zhang 47fdf4416a add room tasks 2025-10-14 09:03:20 +08:00
Yuge Zhang 7f8395b941 Fix Messages Adapter (#150) 2025-10-14 09:01:56 +08:00
Yuge Zhang 3e12d9e774 add variant03 2025-10-14 09:01:29 +08:00
Yuge Zhang badd4de08e . 2025-10-14 01:04:22 +08:00
Yuge Zhang c0c65d5cad fix dataset wrong labels 2025-10-14 01:01:35 +08:00
Yuge Zhang a445b49e9d update data 2025-10-14 00:39:35 +08:00
Yuge Zhang 91b148a966 update all data samples 2025-10-14 00:03:26 +08:00
Yuge Zhang 495d4eba38 Refine algo decorator fallbacks (#146) 2025-10-13 23:46:47 +08:00
Yuge Zhang d6c53cbc52 . 2025-10-13 23:45:18 +08:00
Yuge Zhang 8dd4c5a1a3 Local SFT example and bug fixes (#149) 2025-10-13 23:10:00 +08:00
Yuge Zhang b1ae0b75c4 Ensure Trainer.dev validates FastAlgorithm usage (#148) 2025-10-13 22:02:52 +08:00
Yuge Zhang 31a8097e7d update tests 2025-10-13 21:14:31 +08:00
Yuge Zhang 6c1c53f1ba . 2025-10-13 20:08:14 +08:00
Yuge Zhang c149e66026 . 2025-10-13 19:57:57 +08:00
Yuge Zhang b407430419 apo implementation 2025-10-13 18:51:20 +08:00
Yuge Zhang 9f8ec4950f Refactor emitter module and extract common utilities (#143) 2025-10-12 00:37:44 +08:00
Yuge Zhang c08da2ae37 Use Azure OpenAI and MSR W&B (#142) 2025-10-12 00:01:49 +08:00
Yuge Zhang f032ffa319 Refactor LitAgent decorators and add prompt_rollout support (#141) 2025-10-11 13:56:08 +08:00
Yuge Zhang a4cf2fd5fd Update docs palette and add version warning (#139) 2025-10-10 17:59:01 +00:00
Yuge Zhang 8a4ecbacf6 Improve debugging experience and add new APO example (#138) 2025-10-11 01:42:54 +08:00
Yuge Zhang dd337d456e Add AgentFlow community project links (#137) 2025-10-11 01:28:19 +08:00
Yuge Zhang a0626bdea9 Bird's eye view (v0.2) (#136) 2025-10-10 17:09:02 +08:00
Yuge Zhang 2489d068ba Add env overrides for client/server execution strategy (#135) 2025-10-10 00:43:46 +08:00
Yuge Zhang f4814949cb Supports collecting trace data from LLMProxy (#133) 2025-10-10 00:16:03 +08:00
Yuge Zhang a0791e8b13 Fix flaky pipeline (#131) 2025-10-09 14:54:33 +08:00
Yuge Zhang 26d1df698d Show pytest durations in workflows (#130) 2025-10-09 01:54:06 +08:00
Yuge Zhang 7bf418ea67 Use v0.2 interfaces in Trainer (#129) 2025-10-09 01:53:12 +08:00
Yuge Zhang d735fb27c4 Handle recovering attempts in memory store (#124) 2025-10-09 01:45:50 +08:00
lhx 347638f218 fix: add BaseMessage import to sql_agent.py (#126)
* fix: add BaseMessage import to sql_agent.py

* trigger

---------

Co-authored-by: Yuge Zhang <scottyugochang@gmail.com>
2025-10-09 01:39:41 +08:00
Yuge Zhang a42839b7fb Move test files (#128) 2025-10-08 14:36:54 +08:00
Yuge Zhang e11036cf7b Fix HTTP tracer tests (#127) 2025-10-08 13:30:36 +08:00
Yuge Zhang 685eea70a6 Implement AgentRunnerV2 (#125) 2025-10-03 01:10:09 +08:00
Yuge Zhang a1a4fe39c6 Add missing future import annotations (#123) 2025-10-02 01:48:49 +08:00
Yuge Zhang a9d0c9237d Add LLM proxy (#122) 2025-10-02 01:26:04 +08:00
Yuge Zhang 1513b52a05 Add full pytest workflow on GPU (#121) 2025-10-01 08:06:40 +00:00
xJkie 4ec1029577 Rollout with rollout_manager with REMAX advantage estimator (#115) 2025-10-01 13:10:45 +08:00
Yuge Zhang 63c133051d Add Execution Strategies: Client-Server and Shared Memory implementations (#120) 2025-10-01 13:09:45 +08:00
Yuge Zhang 2316a8451e Add LightningStore interface and implementation (#118) 2025-09-30 09:29:14 +00:00
Yuge Zhang 138ad0e487 Add DeepWerewolf to community projects (#112) 2025-09-27 13:59:56 +08:00
Yuge Zhang 504ef2c627 Add visualization to adapter (#113)
* Add visualization to adapter

* Update agentlightning/adapter/triplet.py

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2025-09-27 12:21:40 +08:00
Yuge Zhang a63197355c Initialize trace adapter (#103) 2025-09-20 12:57:12 +00:00
Yuge Zhang 3eb725fade Update pre-commit configurations and bump to 0.2.0 (#102) 2025-09-20 07:20:14 +00:00
Yuge Zhang 66bcfeba11 Fix pyright issues (#101) 2025-09-20 14:51:56 +08:00
Yuge Zhang a9208ab700 Sort imports (#100) 2025-09-20 11:30:34 +08:00
Yuge Zhang ddc8997b8c Embed algorithm into trainer (#99) 2025-09-19 18:42:58 +08:00
Yuge Zhang 0a92600a4c Ensure blank line after copyright header (#97)
* Ensure copyright header has separating blank line

* fix check script

* trigger pre-commits
2025-09-19 05:32:24 +00:00
Yuge Zhang ba10c845e1 Bug fixes and RL improvements for Agent-framework support (#96) 2025-09-18 21:22:09 +08:00
Yuge Zhang 7ad967daf7 Skip unexpected rollout ids (#85) 2025-09-18 20:43:44 +08:00
Yuge Zhang f6db2dc8ab Migrate to 1ES pipeline (#95) 2025-09-18 20:26:16 +08:00