flow_matching_imagenet32

→FlowMatchingModel

flow_matching_imagenet32(pretrained: bool = False, overrides: object = {})

source edit

Flow Matching velocity field for ImageNet 32 x 32.

Model Size

Parameters

pretrainedbool= False

Reserved for future weight registration; currently a no-op.

**overridesobject= {}

Optional FlowMatchingConfig field overrides.

Returns

FlowMatchingModel

Velocity field configured for ImageNet at 32 x 32.

Notes

Reference: Lipman et al., "Flow Matching for Generative Modeling", ICLR, 2023 (arXiv:2210.02747), Table 1. Reported: 5.02 FID / 3.53 bits per dimension.

Examples

>>> from lucid.models.generative.flow_matching import flow_matching_imagenet32
>>> flow_matching_imagenet32().path
'ot'

Used by 1

lucid.models.generative.flow_matching