Targeted tweaks to DDPMs produce competitive likelihoods and high-quality samples, with learned reverse variances enabling 10x faster sampling and predictable scaling with compute.
For most experiments, we use a batch size of 128, a learning rate of 10−4, and an exponential moving aver- age (EMA) over model parameters with a rate of 0.9999
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2021 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Improved Denoising Diffusion Probabilistic Models
Targeted tweaks to DDPMs produce competitive likelihoods and high-quality samples, with learned reverse variances enabling 10x faster sampling and predictable scaling with compute.