pytorch / pytorch/rl

[Feature Request] Add Support for FQL

Open
#4,433 0 comments 1 reaction 1 assignee View on GitHub

@neuriv is already working on this.

Since Sep 19, 2026.

enhancement
Dominant language
Python
Stars
3.6k
Forks
487
Avg merge
1d 1h
Merged PRs (30d)
207

Description

https://arxiv.org/abs/2502.02538

Add reusable components and a reference implementation for offline and offline-to-online RL.

  • Add a flow-matching policy module with configurable networks and a distilled one-step actor.
  • Add FQLLoss covering flow-matching behavior cloning, policy distillation, Q-maximization, and critic learning.
  • Reuse existing replay buffers, collectors, target updates, and evaluation utilities.
  • Add a state-based SOTA recipe with Hydra configuration, documentation, tests, CI coverage, and benchmark comparisons against the reference implementation.

i'm interested in seeing this in TRL & can make a draft pr for it.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.