playgroundai
/

playground-v2.5-1024px-aesthetic

@@ -13,9 +13,9 @@ inference:
 This repository contains a model that generates highly aesthetic images of resolution 1024x1024, as well as portrait and landscape aspect ratios. You can use the model with Hugging Face 🧨 Diffusers.
-[insert teaser image]
-**Playground v2.5** is a diffusion-based text-to-image generative model, and a successor to Playground v2 [link].
 Playground v2.5 is near state-of-the-art in aesthetic quality. Our user studies demonstrate that our model outperforms SDXL, Playground v2, PIXART-α, DALL-E 3, and Midjourney 5.2.
@@ -45,17 +45,17 @@ The aesthetic quality of Playground v2.5 dramatically outperforms the current st
 Similarly, for multi aspect ratios, we outperform SDXL by a large margin.
-[insert graph for multi aspect ratios]
 Next, we benchmark Playground v2.5 specifically on people-related images, to test Human Preference Alignment. We compared Playground v2.5 against two commonly-used baseline models: SDXL and RealStock v2, a community fine-tune of SDXL that was trained on a realistic people dataset.
 Playground v2.5 outperforms both baselines by a large margin.
-[insert graph for people prompts]
-Lastly, we report metrics using our MJHQ-30K benchmark which we open-sourced with the v2 release. <link> We report both the overall FID and per category FID. All FID metrics are computed at resolution 1024x1024. Our results show that Playground v2.5 outperforms both Playground v2 and SDXL in overall FID and all category FIDs, especially in the people and fashion categories. This is in line with the results of the user study, which indicates a correlation between human preferences and the FID score of the MJHQ-30K benchmark.
-[insert graph for MJHQ-30K benchmark]
 ### How to cite us
 TODO

 This repository contains a model that generates highly aesthetic images of resolution 1024x1024, as well as portrait and landscape aspect ratios. You can use the model with Hugging Face 🧨 Diffusers.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/636c0c4eaae2da3c76b8a9a3/HYUUGfU6SOCHsvyeISQ5Y.png)
+**Playground v2.5** is a diffusion-based text-to-image generative model, and a successor to [Playground v2](https://huggingface.co/playgroundai/playground-v2-1024px-aesthetic).
 Playground v2.5 is near state-of-the-art in aesthetic quality. Our user studies demonstrate that our model outperforms SDXL, Playground v2, PIXART-α, DALL-E 3, and Midjourney 5.2.
 Similarly, for multi aspect ratios, we outperform SDXL by a large margin.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/636c0c4eaae2da3c76b8a9a3/xMB0r-CmR3N6dABFlcV71.png)
 Next, we benchmark Playground v2.5 specifically on people-related images, to test Human Preference Alignment. We compared Playground v2.5 against two commonly-used baseline models: SDXL and RealStock v2, a community fine-tune of SDXL that was trained on a realistic people dataset.
 Playground v2.5 outperforms both baselines by a large margin.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/636c0c4eaae2da3c76b8a9a3/7c-8Stw52OsNtUjse8Slv.png)
+Lastly, we report metrics using our MJHQ-30K benchmark which we [open-sourced](https://huggingface.co/datasets/playgroundai/MJHQ-30K) with the v2 release. We report both the overall FID and per category FID. All FID metrics are computed at resolution 1024x1024. Our results show that Playground v2.5 outperforms both Playground v2 and SDXL in overall FID and all category FIDs, especially in the people and fashion categories. This is in line with the results of the user study, which indicates a correlation between human preferences and the FID score of the MJHQ-30K benchmark.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/636c0c4eaae2da3c76b8a9a3/7tyYDPGUtokh-k18XDSte.png)
 ### How to cite us
 TODO