Visual Text Editing on AnyText benchmark-EN (test)
0.02NEDMiniGPT5
Evaluation Results
| Method | Links | |
|---|---|---|
| MiniGPT5Capability=Text and image generation2024.07 | 0.02 | |
| MM-InterleavedCapability=Text and image generation2024.07 | 0.04 | |
| SEED-LLaMA-14BCapability=Text and image generation2024.07 | 0.11 | |
| TextHarmony*Capability=Text and image generation, Slide-LoRA=false2024.07 | 0.68 | |
| TextHarmonyCapability=Text and image generation, Slide-LoRA=true2024.07 | 0.75 | |
| TextDiffuser-2Capability=Image generation only2024.07 | 0.81 | |
| TextHarmony-GenCapability=Text and image generation, Training mode=image generation/editing data only2024.07 | 0.86 | |
| AnyTextCapability=Image generation only2024.07 | 0.88 |