[pull] main from inclusionAI:main#43
Merged
pull[bot] merged 3 commits intoaxistore80-coder:mainfrom Apr 20, 2026
Merged
Conversation
…acking (#1151) Add KK (Largest Differencing Method) as an alternative to FFD for micro-batch allocation. KK produces more balanced partitions with lower max-min spread, beneficial for RL workloads with variable sequence lengths. Key changes: - Add _KKSet, _KKState, _kk_partition, kk_allocate in seqpack.py - Add packing_algorithm field to MicroBatchSpec (ffd/kk) - Wire KK allocation through dist_rollout and data utils - Add sequence_packing docs (en/zh) and CLI reference updates - Add comprehensive unit tests and torchrun benchmark Refs: #1151
* fix(infra): move data service seed to worker-level config Set random seed once at worker startup instead of per-request during dataset load and epoch reset. This prevents seed re-initialization from interfering with data shuffling across multiple datasets. Key changes: - Add seed field to DataServiceConfig and DataWorkerConfig - Pass seed as CLI arg to worker process, set once in lifespan - Remove seed from WorkerLoadDatasetRequest and _DatasetState - Add datasets_lock for thread-safe dataset load/unload - Update all trainers to pass seed via DataServiceConfig * fix(infra): harden data worker lifecycle concurrency Prevent races between load/unload and stateful endpoints (fetch, reset, save, load) on the data service worker. Key changes: - Add _loading_ids reservation set so load_dataset does not hold datasets_lock across slow I/O (asyncio.to_thread) - Add unloading flag to _DatasetState; unload_dataset drains in-flight state ops via state.lock before dict removal - Introduce _locked_active_state context manager that checks the unloading flag; apply to fetch, reset, save, load - Add 4 deterministic concurrency regression tests covering duplicate-load rejection, unload drain, stale-fetch 409, and cross-dataset non-blocking --------- Co-authored-by: Wentai Zhang <zhangwentai.zwt@antgroup.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )