Class-conditional image generation on ImageNet 512x512 (train val)
2.41FIDStyleGAN-XL
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| StyleGAN-XL2022.12 | 2.41 | 4.06 | 267.75 | 77 | 52 | |
| DiT-XL/2-Gcfg=1.502022.12 | 3.04 | 5.02 | 240.82 | 84 | 54 | |
| DiT-XL/2-GTotal Images(M)=768, Total Gflops=4.03 x 10^11, Classifier-free Guidance=true2023.11 | 3.04 | 5.02 | 240.82 | 84 | 54 | |
| DIFFUSSM-XL-GTotal Images(M)=302, Total Gflops=3.22 x 10^11, Classifier-free Guidance=true2023.11 | 3.41 | 5.84 | 255.06 | 85 | 49 | |
| ADM-G, ADM-U2022.12 | 3.85 | 5.86 | 221.72 | 84 | 53 | |
| ADM-G, ADM-UTotal Images(M)=1385, Total Gflops=4.5 x 10^12, Classifier-free Guidance=true2023.11 | 3.85 | 5.86 | 221.72 | 84 | 53 | |
| U-ViT/2-GTotal Images(M)=512, Total Gflops=6.81 x 10^10, Classifier-free Guidance=true2023.11 | 4.05 | 8.44 | 261.13 | 84 | 48 | |
| DiT-XL/2-Gcfg=1.252022.12 | 4.64 | 5.77 | 174.77 | 81 | 57 | |
| ADM-G2022.12 | 7.72 | 6.57 | 172.71 | 87 | 42 | |
| ADM-GTotal Images(M)=1385, Total Gflops=5.97 x 10^11, Classifier-free Guidance=true2023.11 | 7.72 | 6.57 | 172.71 | 87 | 42 | |
| BigGAN-deep2022.12 | 8.43 | 8.13 | 177.9 | 88 | 29 | |
| ADM-U2022.12 | 9.96 | 5.62 | 121.78 | 75 | 64 | |
| ADM-UTotal Images(M)=1385, Total Gflops=3.9 x 10^12, Classifier-free Guidance=false2023.11 | 9.96 | 5.62 | 121.78 | 75 | 64 | |
| DiT-XL/22022.12 | 12.03 | 7.12 | 105.25 | 75 | 64 | |
| ADM2022.12 | 23.24 | 10.19 | 58.06 | 73 | 60 | |
| ADMTotal Images(M)=1385, Total Gflops=5.97 x 10^11, Classifier-free Guidance=false2023.11 | 23.24 | 10.19 | 58.06 | 73 | 60 |