Text-to-Image Generation on 10-prompt benchmark (full)
0.281V-ErrorLLM Expansion
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LLM ExpansionBase Model=SDXL-Turbo, Inference steps=62026.06 | 0.281 | 0.458 | 0.281 | |
| NaiveBase Model=SDXL-Turbo, Inference steps=62026.06 | 0.292 | 0.465 | 0.296 | |
| EPIGBase Model=SDXL-Turbo, Inference steps=62026.06 | 0.308 | 0.402 | 0.289 | |
| StandardBase Model=SDXL-Turbo, Inference steps=62026.06 | 0.356 | 0.465 | 0.301 |