Update app.py

app.py

@@ -84,7 +84,9 @@ Text-to-image models are notoriously difficult to evaluate. [FID](https://en.wik
 [CLIP Score](https://en.wikipedia.org/wiki/Fr%C3%A9chet_inception_distance) are not enough to accurately state whether a text-to-image model can 
 **generate "good" images**. "Good" is extremely difficult to put into numbers. \n\n
 Instead, the **Open Parti Prompts Leaderboard** uses human feedback from the community to compare images from different text-to-image models to each other.
-
+\n\n
+**Please don't just look at the benchmark but contribute to it 🥺 \n**
+** 👉 Play the [Open Parti Prompts Game](https://huggingface.co/spaces/OpenGenAI/open-parti-prompts) for no more than 3 minutes and to help open-source 🤗.\n**
 """
 
 EXPLANATION = """\n\n
@@ -93,7 +95,7 @@ EXPLANATION = """\n\n
 In more detail, the [Open Parti Prompts Game](https://huggingface.co/spaces/OpenGenAI/open-parti-prompts) collects human preferences that state which generated image 
 best fits a given prompt from the [Parti Prompts](https://huggingface.co/datasets/nateraw/parti-prompts) dataset. Parti Prompts has been designed to challenge
 text-to-image models on prompts of varying categories and difficulty. The images have been pre-generated from the models that are compared in this space.
-For more information of how the images were created, plesae refer to [Open Parti Prompts](https://huggingface.co/spaces/OpenGenAI/open-parti-prompts).
+For more information of how the images were created, please refer to [Open Parti Prompts](https://huggingface.co/spaces/OpenGenAI/open-parti-prompts).
 The community's answers are then stored and used in this space to give a human-preference-based comparison of the different models. \n\n
 
 Currently the leaderboard includes the following models: