Inconsistency and intrinsic difficulty of different versions of text embedding for Wuerstchen. Classification error rate and obtained with Dual Vision Transformer —DaViT— model.
| Training Testing | Wuerstchen | Single prompt | with CLIP tokenizer | with T5xxl | CLIP model with projection |
|---|---|---|---|---|---|
| Wuerstchen | 0.98% | 0.98% | 1.14% | 1.35% | 1.06% |
| Training Testing | Wuerstchen | Single prompt | with CLIP tokenizer | with T5xxl | CLIP model with projection |
|---|---|---|---|---|---|
| Wuerstchen | 0.98% | 0.98% | 1.14% | 1.35% | 1.06% |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.