Update app.py

app.py

@@ -94,6 +94,48 @@ with gr.Blocks(css=css) as demo:
     gr.ChatInterface(
         generate,
         additional_inputs=additional_inputs,
+        markdown="""
+    | Feature | Description | Byline |
+    |---------|-------------|--------|
+    | 🪟 Sliding Window Attention with 128K tokens span | Enables the model to have a larger context for each token. | Increases model's understanding of context, resulting in more coherent and contextually relevant outputs. |
+    | 🚀 GQA for faster inference | Graph Query Attention allows faster computation during inference. | Speeds up the model inference time without sacrificing too much on accuracy. |
+    | 📝 Byte-fallback BPE tokenizer | Uses Byte Pair Encoding but can fall back to byte-level encoding. | Allows the tokenizer to handle a wider variety of input text while keeping token size manageable. |
+    | 📜 License | Released under Apache 2.0 License | Gives you a permissive free software license, allowing you freedom to use, modify, and distribute the code. |
+    | 📦 Usage | | |
+    | 📚 Available on Huggingface Hub | The model can be easily downloaded and set up from Huggingface. | Makes it easier to integrate the model into various projects. |
+    | 🐍 Python code snippets for easy setup | Provides Python code snippets for quick and easy model setup. | Facilitates rapid development and deployment, especially useful for prototyping. |
+    | 📈 Expected speedups with Flash Attention 2 | Upcoming update expected to bring speed improvements. | Keep an eye out for this update to benefit from performance gains. |
+# 🛠 Model Features and More 🛠
+
+## Features
+
+- 🪟 Sliding Window Attention with 128K tokens span  
+  - **Byline**: Increases model's understanding of context, resulting in more coherent and contextually relevant outputs.
+
+- 🚀 GQA for faster inference  
+  - **Byline**: Speeds up the model inference time without sacrificing too much on accuracy.
+
+- 📝 Byte-fallback BPE tokenizer  
+  - **Byline**: Allows the tokenizer to handle a wider variety of input text while keeping token size manageable.
+
+- 📜 License: Released under Apache 2.0 License  
+  - **Byline**: Gives you a permissive free software license, allowing you freedom to use, modify, and distribute the code.
+
+## Usage 📦
+
+- 📚 Available on Huggingface Hub  
+  - **Byline**: Makes it easier to integrate the model into various projects.
+
+- 🐍 Python code snippets for easy setup  
+  - **Byline**: Facilitates rapid development and deployment, especially useful for prototyping.
+
+- 📈 Expected speedups with Flash Attention 2  
+  - **Byline**: Keep an eye out for this update to benefit from performance gains.
+
+        """
+
+        
+        gr.Markdown(markdown)         
         examples = [
             ["🏰 Welcome to the Kingdom of Elandria! You are Jim and Tim, two bumbling bros with a knack for mischief. 🤴🤴 [Action: Introduce yourselves, Equipment: Scepters of Foolishness]"],
             ["🌲 You find yourselves in a forest filled with magical creatures and oddly specific 'Do Not Disturb' signs. 🦄 [Action: Proceed cautiously, Equipment: Map of Social Etiquette]"],