[6/7][multi-lora]: async batch collection and data conversion - per-adapter buffers, round robin collection, batch metadata#1747
Conversation
…ml config, shared helpers, CLI flags and validation
…d, control-plane HTTP API, named Ray actor
…schedules, slot lifecycle, batch routing in get_batch
…ert push to engines, reconcile/train/save hooks
… per-adapter rewards, prefill logprob grouping
…apter buffers, round-robin collection, batch metadata
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a comprehensive multi-LoRA training framework, including per-slot decoupled Adam optimizers, per-adapter learning rate and weight decay schedulers, a background async rollout producer, and a control-plane HTTP API server to manage adapter lifecycles. The review feedback is highly actionable and points out several key areas for improvement: optimizing sequential ray.get calls in loops to run asynchronously, replacing assert statements with ValueError for robust argument validation, guarding against potential runtime errors (such as an IndexError on empty slots or an AttributeError on missing configurations), and avoiding swallowing exceptions silently by logging them instead.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
yushengsu-thu
left a comment
yushengsu-thu
left a comment
There was a problem hiding this comment.
Reviewed extensively on the umbrella PR #1638 (2xH200 E2E validated); layer verified independently (imports + fast suites green on this branch's own tree).
2 participants
#1638