Issues

• Image training example save_model without loss_scaler uses save_checkpoint

  facebookresearch/flow_matching#92 · Filter by author lfrancoisdow opened on May 6, 2026

• Advice on training a language model?

  facebookresearch/flow_matching#89 · Filter by author Mattias421 opened on Oct 8, 2025

• Question about Zero_module

  facebookresearch/flow_matching#87 · Filter by author Chuong98CC opened on Oct 4, 2025

• adding reflow to example

  facebookresearch/flow_matching#86 · Filter by author genji970 opened on Sep 22, 2025
  3

• improve readability

  facebookresearch/flow_matching#85 · Filter by author genji970 opened on Sep 20, 2025

• Question on Riemannian probability path with geodesic interpolation

  facebookresearch/flow_matching#75 · Filter by author johannespitz opened on Jul 23, 2025
  1

• Invalid t shape for data with ndim > 1 in GeodesicProbPath.sample()

  facebookresearch/flow_matching#73 · Filter by author LukasSchweizer opened on Jul 11, 2025

• Implementing Mean Flows for One-Step Sampling

  facebookresearch/flow_matching#65 · Filter by author amorehead opened on Jun 21, 2025
  7

• Compute likelihoods during sampling

  facebookresearch/flow_matching#62 · Filter by author timonpalm opened on Jun 11, 2025

• support for cartesian product maniforlds

  facebookresearch/flow_matching#60 · Filter by author lukasheinrich opened on Jun 6, 2025
  3

• solver.compute_likelihood with enable_grad=True returns a trajectory and logprob whose gradients are set to None

  facebookresearch/flow_matching#59 · Filter by author jbarraquand opened on May 24, 2025
  5

• The scale of timesteps in the text modeling examples

  facebookresearch/flow_matching#17 · Filter by author realfolkcode opened on Dec 16, 2024
  1