Hi-DiT: Hybrid Latent-Pixel Diffusion Transformer for Image Generation

Benchmark Model Rank Results
image-generation-on-imagenet-256x256Hi-DiT (CFG=2.4)–FID: 1.06
image-generation-on-imagenet-512x512Hi-DiT (CFG=4.6)–FID: 1.26Inception score: 295.0