End-to-End Training for Unified Tokenization and Latent Denoising

Benchmark Model Rank Results
image-generation-on-imagenet-256x256UNITE-L#67FID: 1.73
image-generation-on-imagenet-256x256UNITE-XL#68FID: 1.75
image-generation-on-imagenet-256x256UNITE-B#110FID: 2.12