Update app.py

app.py

@@ -393,13 +393,10 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown(
         """
         # The Annotated Discrete Diffusion Models
-        This Gradio demo provides an interactive implementation of the character-level discrete diffusion model.
-        The model starts with random characters (noise) and iteratively denoises them to generate coherent text.
         
-        **How to use:**
-        1. Adjust the **Denoising Steps** and **Replay Speed**.
-        2. Click **Generate**. The model will first compute all steps (see progress bar).
-        3. After computation, the app will replay the denoising process in the textbox below.
+        ## Tiny 7.23M parameters Shakespeare character diffusion model by [Ashwani Kumar](https://x.com/ash_at_tt/status/1977376958859092250)
+        
+        [GitHub](https://github.com/ash80/diffusion-gpt), [Colab](https://colab.research.google.com/github/ash80/diffusion-gpt/blob/master/The_Annotated_Discrete_Diffusion_Models.ipynb)
         """
     )
     
@@ -415,7 +412,7 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
         speed_slider = gr.Slider(
             minimum=1,
             maximum=20,
-            value=5,
+            value=10,
             step=1,
             label="Replay Speed",
             info="Controls the speed of the animation after generation.",
@@ -432,12 +429,8 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
         placeholder="Generation will appear here..."
     )
     
-    # Hidden state to store the diffusion frames
     diffusion_frames_state = gr.State([])
 
-    # Chain the events:
-    # 1. Click button -> run fast generation, store frames in state
-    # 2. After generation -> run slow replay from the state
     generate_event = generate_button.click(
         fn=generate_text, 
         inputs=[steps_slider],