Update app.py

app.py

@@ -1,33 +1,42 @@
 import gradio as gr
-from transformers import pipeline
+from huggingface_hub import InferenceClient
 
-# Initialize the pipeline for code generation and assistance
-pipe = pipeline("text-generation", model="Qwen/Qwen2.5-Coder-32B-Instruct")
+# Function to interact with the model using the Inference API
+def chat_with_model(user_input, hf_api_key):
+    # Initialize the InferenceClient with the provided API key
+    client = InferenceClient(api_key=hf_api_key)
 
-# Function to interact with the model for code-related assistance
-def code_assistance(user_input):
-    # Define the system message to set the context
-    system_message = "You are Qwen, a code assistant created by Alibaba Cloud. You assist with code generation, debugging, and explanation tasks."
-    
-    # Format the prompt with the system message and user input (code-related query)
-    prompt = f"{system_message}\nUser: {user_input}\nAssistant (Code Assistance):"
-    
-    # Use the pipeline to generate the response for code assistance
-    response = pipe(prompt, max_length=512, num_return_sequences=1)
-    
-    # Extract and clean the response to return only the assistant's code suggestion or explanation
-    generated_response = response[0]['generated_text'].split("Assistant (Code Assistance):")[1].strip()
+    # Define the messages for the chat (system message tailored for a code assistant)
+    messages = [
+        {"role": "system", "content": "You are a code assistant that helps with code generation, debugging, and explanations."},
+        {"role": "user", "content": user_input}
+    ]
+
+    # Create a stream for chat completions using the API
+    stream = client.chat.completions.create(
+        model="Qwen/Qwen2.5-Coder-32B-Instruct", 
+        messages=messages, 
+        max_tokens=500,
+        stream=True
+    )
     
-    return generated_response
+    # Collect the generated response from the model
+    response = ""
+    for chunk in stream:
+        response += chunk.choices[0].delta.content
+    return response
 
-# Create the Gradio interface for the code assistance chatbot
+# Create the Gradio interface
 iface = gr.Interface(
-    fn=code_assistance, 
-    inputs=gr.Textbox(lines=5, placeholder="Ask for code help..."), 
+    fn=chat_with_model, 
+    inputs=[
+        gr.Textbox(lines=5, placeholder="Ask me anything about coding..."), 
+        gr.Textbox(lines=1, placeholder="Enter your Hugging Face API key", type="password")  # API key input
+    ], 
     outputs="text", 
-    title="Qwen2.5-Coder Chatbot", 
-    description="A chatbot using Qwen2.5-Coder for code generation, debugging, and explanation tasks."
+    title="Code Assistant with Qwen2.5-Coder", 
+    description="A code assistant that helps you with code generation, debugging, and explanations using the Qwen2.5-Coder model via Hugging Face Inference API."
 )
 
-# Launch the Gradio interface
+# Launch the interface
 iface.launch()