Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design

Jonathan Ho, Xi Chen, Aravind Srinivas, Yan Duan, Pieter Abbeel

Abstract:

Flow-based generative models are powerful exact likelihood models with efficient sampling and inference.
Despite their computational efficiency, flow-based models generally have much worse density modeling performance compared to state-of-the-art autoregressive models. In this paper, we investigate and improve upon three limiting design choices employed by flow-based models in prior work: the use of uniform noise for dequantization, the use of inexpressive affine flows, and the use of purely convolutional conditioning networks in coupling layers. Based on our findings, we propose Flow++, a new flow-based model that is now the state-of-the-art non-autoregressive model for unconditional density estimation on standard image benchmarks. Our work has begun to close the significant performance gap that has so far existed between autoregressive models and flow-based models.

Keywords: Deep Generative Models, Normalizing Flows, RealNVP, Density Estimation

TL;DR: Improved training of current flow-based generative models (Glow and RealNVP) on density estimation benchmarks

Code: [ aravind0706/flowpp](https://github.com/aravind0706/flowpp) + [ 3 community implementations](https://paperswithcode.com/paper/?openreview=Hyg74h05tX)

Data: [CIFAR-10](https://paperswithcode.com/dataset/cifar-10)

Community Implementations: [ 4 code implementations](https://www.catalyzex.com/paper/flow-improving-flow-based-generative-models/code)