Image Generation on ArtBench
18.23FIDSADM (parameter-efficient)
Evaluation results
| Method | Links | |
|---|---|---|
| SADM (parameter-efficient)Pre-training=ImageNet, Resolution=256x256, Tuning Strategy=Parameter-efficient2024.02 | 18.23 | |
| SADM (full)Pre-training=ImageNet, Resolution=256x256, Tuning Strategy=Full Fine-tuning2024.02 | 19.84 | |
| DiffFitPre-training=ImageNet, Resolution=256x256, Tuning Strategy=Parameter-efficient2024.02 | 20.87 | |
| BitFitPre-training=ImageNet, Resolution=256x256, Tuning Strategy=Parameter-efficient2024.02 | 24.53 | |
| Full Fine-tuning with DDPMPre-training=ImageNet, Resolution=256x256, Tuning Strategy=Full Fine-tuning2024.02 | 25.31 | |
| AdaptFormerPre-training=ImageNet, Resolution=256x256, Tuning Strategy=Parameter-efficient2024.02 | 38.43 | |
| VPTPre-training=ImageNet, Resolution=256x256, Tuning Strategy=Parameter-efficient2024.02 | 40.74 | |
| LoRAPre-training=ImageNet, Resolution=256x256, Tuning Strategy=Parameter-efficient2024.02 | 80.99 |