ddpm_imagenet64_gen

→DDPMForImageGeneration

ddpm_imagenet64_gen(pretrained: bool = False, overrides: object = {})

source edit

Construct an Improved-DDPM ImageNet 64x64 model with training loss and .generate().

Same trunk as ddpm_imagenet64 (sample size 64x64, 4-stage U-Net with learn_sigma=True).

Warning:

This factory currently raises NotImplementedError at construction time — the default config has learn_sigma=True, which requires the Improved-DDPM hybrid $L_{\text{simple}} + \lambda L_{\text{vlb}}$ loss not yet implemented in Lucid. Pass learn_sigma=False to override.

Parameters

pretrainedbool= False

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

**overridesobject= {}

Optional DDPMConfig field overrides forwarded into the underlying config.

Returns

DDPMForImageGeneration

Improved-DDPM ImageNet 64x64 wrapper.

Notes

Reference: Nichol and Dhariwal, "Improved Denoising Diffusion Probabilistic Models", ICML, 2021 (arXiv:2102.09672), Table 1.

Examples

>>> import lucid
>>> from lucid.models.generative.ddpm import ddpm_imagenet64_gen
>>> model = ddpm_imagenet64_gen(learn_sigma=False).eval()
>>> x_t = lucid.randn((1, 3, 64, 64))
>>> t = lucid.tensor([1234]).long()
>>> out = model(x_t, t)
>>> out.sample.shape   # (1, 3, 64, 64)
(1, 3, 64, 64)

Used by 1

lucid.models.generative.ddpm

>>> import lucid >>> from lucid.models.generative.ddpm import ddpm_imagenet64_gen >>> model = ddpm_imagenet64_gen(learn_sigma=False).eval() >>> x_t = lucid.randn((1, 3, 64, 64)) >>> t = lucid.tensor([1234]).long() >>> out = model(x_t, t) >>> out.sample.shape # (1, 3, 64, 64) (1, 3, 64, 64)