Adversarial Flow Models

TL;DR

Adversarial flow models combine adversarial training with flow-based generative models, enabling stable, efficient, and high-quality image generation with fewer steps and no intermediate supervision.

Contribution

This work introduces adversarial flow models that learn deterministic mappings, improving stability and efficiency over traditional GANs and flow models, with state-of-the-art results on ImageNet.

Findings

Achieved a new best FID of 2.38 on ImageNet-256px.

End-to-end training of very deep models with high performance.

Surpassed shallower models with similar compute and parameters.

Abstract

We present adversarial flow models, a class of generative models that belongs to both the adversarial and flow families. Our method supports native one-step and multi-step generation and is trained with an adversarial objective. Unlike traditional GANs, in which the generator learns an arbitrary transport map between the noise and data distributions, our generator is encouraged to learn a deterministic noise-to-data mapping. This significantly stabilizes adversarial training. Unlike consistency-based methods, our model directly learns one-step or few-step generation without having to learn the intermediate timesteps of the probability flow for propagation. This preserves model capacity and avoids error accumulation. Under the same 1NFE setting on ImageNet-256px, our B/2 model approaches the performance of consistency-based XL/2 models, while our XL/2 model achieves a new best FID of 2.38. We additionally demonstrate end-to-end training of 56-layer and 112-layer models without any intermediate supervision, achieving FIDs of 2.08 and 1.94 with a single forward pass and surpassing the corresponding 28-layer 2NFE and 4NFE counterparts with equal compute and parameters. The code is available at https://github.com/ByteDance-Seed/Adversarial-Flow-Models