Update app.py

app.py

@@ -1,70 +1,101 @@
-import gradio as gr
-from huggingface_hub import InferenceClient
+'''
+This script creates a Gradio chatbot interface for the ibm-granite/granite-3.3-8b-instruct model.
 
+Key Features:
+- Loads the model and tokenizer from Hugging Face Hub.
+- Uses a chat interface for interactive conversations.
+- Manages chat history to maintain context.
+- Handles API key management through Hugging Face Spaces secrets.
+'''
 
-def respond(
-    message,
-    history: list[dict[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-    hf_token: gr.OAuthToken,
-):
-    """
-    For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-    """
-    client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
+import gradio as gr
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer, set_seed
+import os
 
-    messages = [{"role": "system", "content": system_message}]
+# --- Configuration ---
+MODEL_ID = "ibm-granite/granite-3.3-8b-instruct"
 
-    messages.extend(history)
+# --- Model and Tokenizer Loading ---
+def load_model_and_tokenizer():
+    '''Load the model and tokenizer, handling potential errors.'''
+    try:
+        # Securely get the Hugging Face token from secrets
+        hf_token = os.getenv("HUGGINGFACE_TOKEN")
+        if not hf_token:
+            raise ValueError("HUGGINGFACE_TOKEN secret not found. Please add it to your Space settings.")
 
-    messages.append({"role": "user", "content": message})
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        
+        model = AutoModelForCausalLM.from_pretrained(
+            MODEL_ID,
+            device_map=device,
+            torch_dtype=torch.bfloat16,
+            token=hf_token
+        )
+        tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, token=hf_token)
+        
+        return model, tokenizer, device
+    except Exception as e:
+        # Provide a user-friendly error message
+        raise RuntimeError(f"Failed to load model or tokenizer: {e}")
 
-    response = ""
+model, tokenizer, device = load_model_and_tokenizer()
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        choices = message.choices
-        token = ""
-        if len(choices) and choices[0].delta.content:
-            token = choices[0].delta.content
+# --- Chatbot Logic ---
+def chat_function(message, history):
+    '''
+    This function processes the user's message and returns the model's response.
+    '''
+    # Set seed for reproducibility
+    set_seed(42)
 
-        response += token
-        yield response
+    # Format the conversation history for the model
+    conv = []
+    for user_msg, model_msg in history:
+        conv.append({"role": "user", "content": user_msg})
+        conv.append({"role": "assistant", "content": model_msg})
+    conv.append({"role": "user", "content": message})
 
+    # Tokenize the input
+    input_ids = tokenizer.apply_chat_template(
+        conv, 
+        return_tensors="pt", 
+        thinking=False,  # Set to False for direct response
+        add_generation_prompt=True
+    ).to(device)
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-chatbot = gr.ChatInterface(
-    respond,
-    type="messages",
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
+    # Generate the response
+    output = model.generate(
+        input_ids,
+        max_new_tokens=1024,
+        do_sample=True,
+        top_k=50,
+        top_p=0.95,
+        temperature=0.7,
+    )
 
-with gr.Blocks() as demo:
-    with gr.Sidebar():
-        gr.LoginButton()
-    chatbot.render()
+    # Decode the prediction
+    prediction = tokenizer.decode(output[0, input_ids.shape[1]:], skip_special_tokens=True)
+    
+    return prediction
 
+# --- Gradio Interface ---
+def create_gradio_interface():
+    '''Create and return the Gradio ChatInterface.'''
+    return gr.ChatInterface(
+        fn=chat_function,
+        title="Granite 3.3 8B Chatbot",
+        description="A chatbot powered by the ibm-granite/granite-3.3-8b-instruct model. Ask any question!",
+        theme="soft",
+        examples=[
+            ["Hello, who are you?"],
+            ["What is the capital of France?"],
+            ["Explain the theory of relativity in simple terms."]
+        ]
+    )
 
+# --- Main Execution ---
 if __name__ == "__main__":
-    demo.launch()
+    chatbot_interface = create_gradio_interface()
+    chatbot_interface.launch()