vae_gen

→VAEForImageGeneration

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

source edit

Construct a vanilla VAE with the ELBO loss and prior-sample .generate().

Same trunk as vae (single 128-dim bottleneck, 3-stage encoder), wrapped with the ELBO training loss (reconstruction + $\beta$ -weighted KL) and a convenience prior sampler.

Model Size

Parameters

pretrainedbool= False

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

**overridesobject= {}

Optional VAEConfig field overrides forwarded into the underlying config. Pass kl_weight=... to engage

\beta

-VAE behaviour; pass recon_loss="bce" for Bernoulli likelihoods.

Returns

VAEForImageGeneration

Vanilla VAE wrapped with ELBO loss and prior sampler.

Notes

Reference: Kingma and Welling, "Auto-Encoding Variational Bayes", ICLR, 2014 (arXiv:1312.6114).

Examples

>>> import lucid
>>> from lucid.models.generative.vae import vae_gen
>>> model = vae_gen().eval()
>>> x = lucid.randn((1, 3, 32, 32))
>>> out = model(x)
>>> out.sample.shape, out.kl_loss.shape   # reconstruction + scalar KL
((1, 3, 32, 32), ())

Used by 1

lucid.models.generative.vae

>>> import lucid >>> from lucid.models.generative.vae import vae_gen >>> model = vae_gen().eval() >>> x = lucid.randn((1, 3, 32, 32)) >>> out = model(x) >>> out.sample.shape, out.kl_loss.shape # reconstruction + scalar KL ((1, 3, 32, 32), ())