Repository navigation
Conversation
Shared LoRA pools are always routed and grow to the largest sampler_replicas any job asks for. The dispatcher sends round-robin by job, only FFT pools switch revisions and do so once the current revision drains, and the API server recreates a missing pool for a live job or fails its requests. Adds the open item that a pool's replicas can land on different GPU types.
ShubyM
marked this pull request as ready for review
October 26, 2026 17:35
This branch has not been deployed
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 join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Addresses Hoda's comments on the previous design (#324) and matches the doc to the code in #357 and #378, plus the fairness and failures PRs that follow.
sampler_replicas, the pool grows to the largest and never shrinks. Placement is best effort, and node role labels keep sampler growth off trainer GPUs.New open item: a pool's replicas can land on different GPU types, since placement asks only for memory. Pinning later replicas to the first replica's product is a follow-up.