Update README.md
Browse files
README.md
CHANGED
|
@@ -89,6 +89,15 @@ We then presented 30 side by side comparisons to a group of professionals, who v
|
|
| 89 |
|
| 90 |
According to the results, DeciDiffusion at 30 iterations exhibits an edge in aesthetics, but when it comes to prompt alignment, it’s on par with Stable Diffusion at 50 iterations.
|
| 91 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
## Runtime Benchmarks
|
| 93 |
|
| 94 |
The following tables provide an image latency comparison between DeciDiffusion 1.0 and Stable Diffusion 1.5.
|
|
|
|
| 89 |
|
| 90 |
According to the results, DeciDiffusion at 30 iterations exhibits an edge in aesthetics, but when it comes to prompt alignment, it’s on par with Stable Diffusion at 50 iterations.
|
| 91 |
|
| 92 |
+
The following table summarizes our survey results:
|
| 93 |
+
|
| 94 |
+
|Answer| Better image aesthetics | Better prompt alignment |
|
| 95 |
+
|:----------|:----------|:----------|
|
| 96 |
+
| DeciDiffusion 30 Iterations | 41.4% | 20.8% |
|
| 97 |
+
| StableDiffusion 50 Iterations | 30.5% |18.8% |
|
| 98 |
+
| On Par | 26.3% |39.1% |
|
| 99 |
+
| Neither | 2.1% | 11.4%|
|
| 100 |
+
|
| 101 |
## Runtime Benchmarks
|
| 102 |
|
| 103 |
The following tables provide an image latency comparison between DeciDiffusion 1.0 and Stable Diffusion 1.5.
|