flow-matching

library1.0.10pypypiunverified

Flow Matching for Generative Modeling, a library by Meta Research for training continuous normalizing flows via flow matching objectives. Current version 1.0.10, supports Python >=3.9, actively maintained.

ai-ml

pip install flow-matching

github ↗package ↗

INSTALL

IMPORT

SIG · FLOW-MATCHING

flow-matching

ai-mlpythonv1.0.10

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

FlowMatching

✓ from flow_matching import FlowMatching

✗ from flow_matching import FlowMatching

Minimal example: create a flow matching model, compute the flow matching objective between random source and target samples.

import torch from flow_matching import CondOTFlowMatching from flow_matching.objectives import FMObjective dim = 2 flow = CondOTFlowMatching(dim=dim) objective = FMObjective(flow) x0 = torch.randn(100, dim) x1 = torch.randn(100, dim) t = torch.rand(100) loss = objective(x0, x1, t) print(f"Loss: {loss.item()}")

Debug

Known issues

breakingAs of v1.0.0, the API was completely rewritten. Imports like `from flow_matching import FlowMatching` now return a high-level wrapper, not the old `CNF` class. If upgrading from <1.0, review all imports.

fix

Update imports: use `from flow_matching import CondOTFlowMatching` instead of old `from flow_matching import FlowMatching`.

affects: <1.0.0

deprecatedThe `models.CNF` class is deprecated since v1.0.0. Use `CondOTFlowMatching` or other specific flow matching classes.

fix

Replace `from flow_matching.models import CNF` with `from flow_matching import CondOTFlowMatching`.

affects: >=1.0.0

gotchaThe library expects input tensors (x0, x1) to have dtype `torch.float32`. Using float64 may cause silent errors or performance degradation.

fix

Ensure tensors are float32: `x0 = x0.float()`.

affects: all

Upgrade

Version history

1.0.10latest on PyPI · released Dec 18, 2024

Audit

Dependencies

torchrequiredCore dependency for tensor operations and neural networks.

torchdiffeqrequiredUsed for ODE solvers in sampling.

Agent activity

17 hits · last 30 days

node

OpenAI (training)

Resources

githubhttps://github.com/facebookresearch/flow_matching ↗

packagehttps://pypi.org/project/flow-matching/ ↗