GitHub · Apple Machine Learning
Recent efforts such as TARFlow have shown that NFs are capable of achieving promising performance on image modeling tasks
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
★ Tier-1 Source
In this work, Apple further advance the state of Normalizing Flow generative models by introducing iterative TARFlow (iTARFlow).
Key facts
- Through extensive experiments, they show that iTARFlow achieves competitive performance across ImageNet resolutions of 64, 128, and 256 pixels, demonstrating its potential as a strong generative model
- Authors Tianrong Chen, Jiatao Gu, David Berthelot, Joshua Susskind, Shuangfei Zhai
- Normalizing Flows (NFs) are a classical family of likelihood-based methods that have received revived attention
- In this work, they further advance the state of Normalizing Flow generative models by introducing iterative TARFlow (iTARFlow)
Summary
Authors Tianrong Chen, Jiatao Gu, David Berthelot, Joshua Susskind, Shuangfei Zhai. Normalizing Flows (NFs) are a classical family of likelihood-based methods that have received revived attention. Through extensive experiments, they show that iTARFlow achieves competitive performance across ImageNet resolutions of 64, 128, and 256 pixels, demonstrating its potential as a strong generative model and advancing the frontier of Normalizing Flows.