Update app.py

app.py

@@ -23,7 +23,15 @@ def respond(
     seed,
     custom_model
 ):
+    """
+    This function handles the conversation logic and streams the response.
 
+    Arguments:
+    - message: The new user message
+    - history: Chat history in the form of a list of (user_message, assistant_message) pairs
+    - system_message: The system prompt specifying how the assistant should behave
+    - max_tokens, temperature, top_p, frequency_penalty, seed, custom_model: Various parameters for text generation
+    """
     print(f"Received message: {message}")
     print(f"History: {history}")
     print(f"System message: {system_message}")
@@ -35,6 +43,7 @@ def respond(
     if seed == -1:
         seed = None
 
+    # Create the base system-level message
     messages = [{"role": "system", "content": system_message}]
     print("Initial messages array constructed.")
 
@@ -61,6 +70,7 @@ def respond(
     response = ""
     print("Sending request to OpenAI API.")
 
+    # Stream tokens from the HF inference endpoint
     for message_chunk in client.chat.completions.create(
         model=model_to_use,
         max_tokens=max_tokens,
@@ -78,13 +88,29 @@ def respond(
 
     print("Completed response generation.")
 
-# GRADIO UI
 
-chatbot = gr.Chatbot(height=600, show_copy_button=True, placeholder="Select a model and begin chatting", likeable=True, layout="panel")
+# -------------------------
+# Gradio UI definitions
+# -------------------------
+
+# Chatbot interface
+chatbot = gr.Chatbot(
+    height=600, 
+    show_copy_button=True, 
+    placeholder="Select a model and begin chatting", 
+    likeable=True, 
+    layout="panel"
+)
 print("Chatbot interface created.")
 
-system_message_box = gr.Textbox(value="", placeholder="You are a helpful assistant.", label="System Prompt")
+# System prompt textbox
+system_message_box = gr.Textbox(
+    value="", 
+    placeholder="You are a helpful assistant.", 
+    label="System Prompt"
+)
 
+# Sliders
 max_tokens_slider = gr.Slider(
     minimum=1,
     maximum=4096,
@@ -121,7 +147,9 @@ seed_slider = gr.Slider(
     label="Seed (-1 for random)"
 )
 
-# The custom_model_box is what the respond function sees as "custom_model"
+# This textbox is what the respond() function sees as "custom_model"
+# We will visually place it inside the Model Selection accordion (below),
+# but we define it here so it can be passed to the ChatInterface.
 custom_model_box = gr.Textbox(
     value="",
     label="Custom Model",
@@ -129,14 +157,7 @@ custom_model_box = gr.Textbox(
     placeholder="meta-llama/Llama-3.3-70B-Instruct"
 )
 
-def set_custom_model_from_radio(selected):
-    """
-    This function will get triggered whenever someone picks a model from the 'Featured Models' radio.
-    We will update the Custom Model text box with that selection automatically.
-    """
-    print(f"Featured model selected: {selected}")
-    return selected
-
+# Create the ChatInterface, referencing the respond function and including all inputs
 demo = gr.ChatInterface(
     fn=respond,
     additional_inputs=[
@@ -146,7 +167,7 @@ demo = gr.ChatInterface(
         top_p_slider,
         frequency_penalty_slider,
         seed_slider,
-        custom_model_box,
+        custom_model_box,  # We pass it here to the ChatInterface function
     ],
     fill_height=True,
     chatbot=chatbot,
@@ -154,36 +175,67 @@ demo = gr.ChatInterface(
 )
 print("ChatInterface object created.")
 
+
+# --------------------------
+# Additional Model Selection
+# --------------------------
+
+# This is the function that updates the Custom Model textbox whenever the user picks a model from the Radio
+def set_custom_model_from_radio(selected):
+    """
+    Triggered when the user picks a model from the 'Featured Models' radio.
+    We will update the Custom Model text box with that selection automatically.
+    """
+    print(f"Featured model selected: {selected}")
+    return selected
+
+# The set of models displayed in the radio
+models_list = [
+    "meta-llama/Llama-3.3-70B-Instruct",
+    "meta-llama/Llama-3.2-3B-Instruct",
+    "meta-llama/Llama-3.2-1B-Instruct",
+    "meta-llama/Llama-3.1-8B-Instruct",
+    "NousResearch/Hermes-3-Llama-3.1-8B",
+    "google/gemma-2-27b-it",
+    "google/gemma-2-9b-it",
+    "google/gemma-2-2b-it",
+    "mistralai/Mistral-Nemo-Instruct-2407",
+    "mistralai/Mixtral-8x7B-Instruct-v0.1",
+    "mistralai/Mistral-7B-Instruct-v0.3",
+    "Qwen/Qwen2.5-72B-Instruct",
+    "Qwen/QwQ-32B-Preview",
+    "PowerInfer/SmallThinker-3B-Preview",
+    "HuggingFaceTB/SmolLM2-1.7B-Instruct",
+    "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
+    "microsoft/Phi-3.5-mini-instruct",
+]
+print("Models list initialized.")
+
+# This function handles searching for models by a user-provided filter
+def filter_models(search_term):
+    print(f"Filtering models with search term: {search_term}")
+    filtered = [m for m in models_list if search_term.lower() in m.lower()]
+    print(f"Filtered models: {filtered}")
+    return gr.update(choices=filtered)
+
+
+# --------------------------------
+# Advanced UI arrangement with demo
+# --------------------------------
 with demo:
+    # Create an Accordion for model selection
     with gr.Accordion("Model Selection", open=False):
-        model_search_box = gr.Textbox(
-            label="Filter Models",
-            placeholder="Search for a featured model...",
-            lines=1
-        )
-        print("Model search box created.")
-
-        models_list = [
-            "meta-llama/Llama-3.3-70B-Instruct",
-            "meta-llama/Llama-3.2-3B-Instruct",
-            "meta-llama/Llama-3.2-1B-Instruct",
-            "meta-llama/Llama-3.1-8B-Instruct",
-            "NousResearch/Hermes-3-Llama-3.1-8B",
-            "google/gemma-2-27b-it",
-            "google/gemma-2-9b-it",
-            "google/gemma-2-2b-it",
-            "mistralai/Mistral-Nemo-Instruct-2407",
-            "mistralai/Mixtral-8x7B-Instruct-v0.1",
-            "mistralai/Mistral-7B-Instruct-v0.3",
-            "Qwen/Qwen2.5-72B-Instruct",
-            "Qwen/QwQ-32B-Preview",
-            "PowerInfer/SmallThinker-3B-Preview",
-            "HuggingFaceTB/SmolLM2-1.7B-Instruct",
-            "TinyLlama/TinyLlama-1.1B-Chat-v1.0",
-            "microsoft/Phi-3.5-mini-instruct",
-        ]
-        print("Models list initialized.")
+        # Place the Filter Models textbox and the Custom Model textbox side by side
+        with gr.Row():
+            model_search_box = gr.Textbox(
+                label="Filter Models",
+                placeholder="Search for a featured model...",
+                lines=1
+            )
+            # Render the already-defined 'custom_model_box' so it appears in this row
+            custom_model_box.render()
 
+        # Create the Radio for featured models
         featured_model_radio = gr.Radio(
             label="Select a model below",
             choices=models_list,
@@ -192,12 +244,7 @@ with demo:
         )
         print("Featured models radio button created.")
 
-        def filter_models(search_term):
-            print(f"Filtering models with search term: {search_term}")
-            filtered = [m for m in models_list if search_term.lower() in m.lower()]
-            print(f"Filtered models: {filtered}")
-            return gr.update(choices=filtered)
-
+        # Link the search box to the filtering function
         model_search_box.change(
             fn=filter_models,
             inputs=model_search_box,
@@ -205,6 +252,7 @@ with demo:
         )
         print("Model search box change event linked.")
 
+        # Link the radio to the function that sets the custom model textbox
         featured_model_radio.change(
             fn=set_custom_model_from_radio,
             inputs=featured_model_radio,
@@ -214,6 +262,10 @@ with demo:
 
 print("Gradio interface initialized.")
 
+
+# -----------------------
+# Launch the application
+# -----------------------
 if __name__ == "__main__":
     print("Launching the demo application.")
     demo.launch()