Update app.py

app.py

@@ -1,64 +1,92 @@
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
+from transformers import pipeline
 
-"""
-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("HuggingFaceH4/zephyr-7b-beta")
+# Initialize the pipeline with settings that work well on Hugging Face
+def initialize_model():
+    # HF-specific configuration for memory efficiency
+    pipe = pipeline(
+        "text-generation",
+        model="abhinand/tamil-llama-7b-instruct-v0.2",
+        device_map="auto",
+        torch_dtype="auto",
+        model_kwargs={"load_in_8bit": True}  # 8-bit quantization for HF Spaces
+    )
+    return pipe
 
+# Generate response
+def generate_response(pipe, user_input, chat_history):
+    # Format messages for the model
+    messages = []
+    for human, bot in chat_history:
+        messages.append({"role": "user", "content": human})
+        messages.append({"role": "assistant", "content": bot})
+    
+    # Add the current message
+    messages.append({"role": "user", "content": user_input})
+    
+    try:
+        # Generate response with settings suitable for Spaces
+        response = pipe(
+            messages,
+            max_length=256,  # Shorter responses to save compute
+            do_sample=True,
+            temperature=0.7,
+            top_p=0.9,
+            num_return_sequences=1
+        )
+        
+        # Extract the generated text
+        generated_text = response[0]['generated_text']
+        
+        # Extract only the assistant's response
+        for msg in generated_text:
+            if isinstance(msg, dict) and msg.get("role") == "assistant":
+                return msg.get("content", "")
+        
+        # Fallback if no assistant response is found
+        if isinstance(generated_text, str):
+            return generated_text
+        return "சரியான பதிலைக் கண்டுபிடிக்க முடியவில்லை." # Could not find proper response
+        
+    except Exception as e:
+        print(f"Error generating response: {e}")
+        return f"பிழை ஏற்பட்டது. மீண்டும் முயற்சிக்கவும்." # Error occurred, please try again
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    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)",
-        ),
-    ],
-)
+# Create the Gradio interface
+def create_chatbot_interface():
+    with gr.Blocks() as demo:
+        gr.Markdown("# தமிழ் உரையாடல் பொத்தான் (Tamil Chatbot)")
+        
+        chatbot = gr.Chatbot(label="உரையாடல் (Conversation)")
+        msg = gr.Textbox(label="உங்கள் செய்தி (Your Message)", placeholder="இங்கே தட்டச்சு செய்யவும்...")
+        clear = gr.Button("அழி (Clear)")
+        
+        # Initialize model only once when first needed
+        model = gr.State(None)
+        
+        def load_model_if_needed(model_state):
+            if model_state is None:
+                return initialize_model()
+            return model_state
+        
+        def respond(message, chat_history, model_state):
+            # Load model if not already loaded
+            if model_state is None:
+                model_state = initialize_model()
+                
+            bot_message = generate_response(model_state, message, chat_history)
+            chat_history.append((message, bot_message))
+            return "", chat_history, model_state
+        
+        msg.submit(respond, [msg, chatbot, model], [msg, chatbot, model])
+        clear.click(lambda: None, None, chatbot, queue=False)
+    
+    return demo
 
+# Create and launch the demo
+demo = create_chatbot_interface()
 
+# This is the key part for Hugging Face Spaces
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()