Search results

Image training example save_model without loss_scaler uses save_checkpoint

Status: Open.

#92 In facebookresearch/flow_matching;

· lfrancoisdow opened on May 6, 2026

Advice on training a language model?

Status: Open.

#89 In facebookresearch/flow_matching;

· Mattias421 opened on Oct 8, 2025

Question about Zero_module

Status: Open.

#87 In facebookresearch/flow_matching;

· Chuong98CC opened on Oct 4, 2025

adding reflow to example

Status: Open.

#86 In facebookresearch/flow_matching;

· genji970 opened on Sep 22, 2025

improve readability

Status: Open.

#85 In facebookresearch/flow_matching;

· genji970 opened on Sep 20, 2025

Question on Riemannian probability path with geodesic interpolation

Status: Open.

#75 In facebookresearch/flow_matching;

· johannespitz opened on Jul 23, 2025

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

Status: Open.

#73 In facebookresearch/flow_matching;

· LukasSchweizer opened on Jul 11, 2025

Resuming training from a checkpoint: Incorrect function call and misspelt attribute in ckpt loading logic

Status: Open.

#71 In facebookresearch/flow_matching;

· DhruvaRajwade opened on Jul 8, 2025

Implementing Mean Flows for One-Step Sampling

Status: Open.

#65 In facebookresearch/flow_matching;

· amorehead opened on Jun 21, 2025

Compute likelihoods during sampling

Status: Open.

#62 In facebookresearch/flow_matching;

· timonpalm opened on Jun 11, 2025

support for cartesian product maniforlds

Status: Open.

#60 In facebookresearch/flow_matching;

· lukasheinrich opened on Jun 6, 2025

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

Status: Open.

#59 In facebookresearch/flow_matching;

· jbarraquand opened on May 24, 2025