DiGIT

Reported on 4 benchmarks across 2 tasks · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision2 results

Conditional Image GenerationonImageNet 256x256
FID· 2024-10-16
3.39
SOTA
Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective arXiv:2410.12490
Conditional Image GenerationonImageNet 256x256
Inception score· 2024-10-16
205.96
best: 221.57 (ADM-G + EDS + ECT (ED-DPM, classifier_scale=1.0))
Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective arXiv:2410.12490

Medical2 results

Image GenerationonImageNet 256x256
FID· 2024-10-16
3.39
best: 1.06 (SiT-XL/2 + UCGM-S (E2E-VAE + 40 sampling steps + CFG))
Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective arXiv:2410.12490
Image GenerationonImageNet 256x256
Inception score· 2024-10-16
205.96
best: 338.08 (LEGO-XL)
Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective arXiv:2410.12490