Update support_docs.py

support_docs.py

@@ -34,24 +34,26 @@ def create_support_docs():
         with gr.Accordion("📝 Step 1: Configure Your Space", open=False):
             gr.Markdown("""
             ### Configuration Tab Sections
+    
+            #### 1a. Templates & Identity
+            Included are several pre-populated templates to get you started with playful (but applied) configurations. Or start with the "None (Custom)" option for a blank slate. Personalize the chat u/i and giving it a name, header (max 60 char) and a markdown description for the `README.md` file, then select from Gradio themes like Default, Soft, Glass, Monochrome, or Base.
+    
+            #### 1b. Model Configuration
+            Craft a system prompt that steers your model's behavior using a well-defined role, context and audience, code of conduct, and/or terms of engagement——think of guardrails as creative constraints. Afterward select a model ID from the menu of options, then experiment with sampling parameters, adjusting temperature (values from 0 to 2) to control response variability and/or setting a token response limit (between 50-4096) to cap output.
+    
+            #### 1c. Example Prompts
+            Add 3 to 5 starter prompts that showcase the purpose and design of your space. These examples appear as quick-start buttons for users to try and meant to help them get started.
+    
+            #### 1d. URL Grounding
+            Add up to 10 reference URLs that ground the model in the context of different webpages. The first 2 URLs function as primary sources that will always be loaded with each query, while other URLs serve as supplementary context. Note that only 4096 total characters from URLs are appended to the system prompt.
+    
+            #### 1e. API Key and HuggingFace Token
             
-            **1a. Templates & Identity** Included are several pre-populated templates to get you started with playful (but applied) configurations. Or start with the "None (Custom)" option for a blank slate. Personalize the chat u/i and giving it a name, header (max 60 char) and a markdown description for the `README.md` file, then select from Gradio themes like Default, Soft, Glass, Monochrome, or Base.
-
-            **1b. Model Configuration**: Craft a system prompt that steers your model's behavior using a well-defined role, context and audience, code of conduct, and/or terms of engagement——think of guardrails as creative constraints. Afterward select a model ID from the menu of options, then experiment with sampling parameters, adjusting temperature (values from 0 to 2) to control response variability and/or setting a token response limit (between 50-4096) to cap output.
-
-            **1c. Example Prompts**: Add 3 to 5 starter prompts that showcase the purpose and design of your space. These examples appear as quick-start buttons for users to try and meant to help them get started.
-
-            **1d. URL Grounding**: Add up to 10 reference URLs that ground the model in the context of different webpages. The first 2 URLs function as primary sources that will always be loaded with each query, while other URLs serve as supplementary context. Note that only 4096 total characters from URLs are appended to the system prompt.
-
-            **1e. API Key and HuggingFace Token**
-
-            - Required: Provide the required OpenAI-standard API key (`sk-`) and name for its variable (`API_KEY`) to call the model through a provider like OpenRouter
-                        
-            - Recommended: Add a HuggingFace token (`HF_TOKEN`) to make your space customizable over time through a configuration editor 
-                        
-            - Optional: Set an optional password (`ACCESS_CODE`) to limit who can access your space. Any value goes.
-            
-            **1f. Upload `config.json`**: Done this before? Upload an existing `config.json` files to redeploy your Space with new settings and recent versions.
+            - **Required**: Provide the required OpenAI-standard API key (`sk-`) and name for its variable (`API_KEY`) to call the model through a provider like OpenRouter.
+            - **Recommended**: Add a HuggingFace token (`HF_TOKEN`) to make your space customizable over time through a configuration editor. 
+            - **Optional**: Set an optional password (`ACCESS_CODE`) to limit who can access your space. Any value goes.
+    
+            #### 1f. Upload `config.json`Done this before? Upload an existing `config.json` files to redeploy your Space with new settings and recent versions.
             """)
 
         with gr.Accordion("💬 Step 2: Preview Your Assistant", open=False):