Update README.md

models/kolors/README.md

@@ -5,7 +5,7 @@
 
 We used several typical prompts to visualize generated images, proving that the compilation acceleration of onediff is almost lossless.
 
-| Prompt | Reference Image | OnFlow Backend | NexFort Backend |
+| Prompt | Reference Image (Original PyTorch) | OneFlow Backend | NexFort Backend |
 |--------|-----------------|----------------|-----------------|
 | 穿着黑色T恤的可爱小狗，T恤上面中文绿色大字写着“可图”。 | <img src="./asset/kolors_01.png" width="150px"> | <img src="./asset/kolors_oneflow_compile_01.png" width="150px"> | <img src="./asset/kolors_nexfort_compile_01.png" width="150px"> |
 | 一张兔子的特写照片，春天的森林中，有雾、晕影、开花、戏剧性的氛围，以三分法为中心，200mm 1.4f的微距镜头拍摄。 | <img src="./asset/kolors_02.png" width="150px"> | <img src="./asset/kolors_oneflow_compile_02.png" width="150px"> | <img src="./asset/kolors_nexfort_compile_02.png" width="150px"> |
@@ -16,3 +16,22 @@ We used several typical prompts to visualize generated images, proving that the
 
 
 ## Quantitative evaluation
+
+
+We calculated the Frechet Inception Distance (FID), Inception Score (IS), CLIP Score, and [Aesthetic Score](https://github.com/christophschuhmann/improved-aesthetic-predictor) based on the [coco-30-val-2014](https://huggingface.co/datasets/sayakpaul/coco-30-val-2014) dataset.
+
+The Human [Preference Score v2 (HPS v2)](https://github.com/tgxs002/HPSv2) was evaluated on the Human Preference Dataset v2 (HPD v2) dataset.
+
+These metrics demonstrate that the acceleration by onediff does not result in quality loss.
+
+In addition, compared with the images generated by the original pytorch HF diffusers baseline, we calculated the average Structural Similarity (SSIM) to demonstrate that the structure of the images generated after acceleration by onediff does not undergo significant changes.
+
+
+| Metric            | Reference Value (Original PyTorch) | OneFlow Backend | NexFort Backend |
+|-------------------|------------------------------------|-----------------|-----------------|
+| FID               | 25.022                             | 24.924          | 24.904          |
+| IS                | 37.472 ± 0.474                     | 37.575 ± 0.559  | 37.415 ± 0.723  |
+| CLIP Score        | -                                  | -               | -               |
+| Aesthetic Score   | 6.098                              | 6.097           | 6.100           |
+| HPS v2            | -                                  | -               | -               |
+| SSIM              | -                                  | -               | -               |