Add app.py

app.py

@@ -0,0 +1,55 @@
+import gradio as gr
+from llama_cpp import Llama
+import os
+
+MODEL_PATH = "llama-2-7b-chat.ggmlv3.q2_K.bin"
+
+# Load model at startup
+print("Loading Llama-2 model...")
+llm = None
+if os.path.exists(MODEL_PATH):
+    llm = Llama(model_path=MODEL_PATH, n_ctx=2048, n_threads=4, verbose=False)
+    print("Model loaded!")
+else:
+    print(f"Model not found at {MODEL_PATH}. Upload the model file to the Space.")
+
+def chat(message, history):
+    if llm is None:
+        return "Model not loaded. Please upload 'llama-2-7b-chat.ggmlv3.q2_K.bin' to this Space."
+    
+    # Build conversation context from history
+    context = ""
+    for user_msg, bot_msg in history[-5:]:  # use last 5 turns
+        context += f"[INST] {user_msg} [/INST] {bot_msg} </s>"
+    
+    prompt = f"[INST] <<SYS>>\nYou are a helpful AI assistant.\n<</SYS>>\n\n{context}[INST] {message} [/INST]"
+    
+    output = llm(prompt, max_tokens=512, stop=["[/INST]", "</s>"], echo=False)
+    response = output["choices"][0]["text"].strip()
+    return response
+
+# Gradio chat interface
+demo = gr.ChatInterface(
+    fn=chat,
+    title="Llama-2-7B Chatbot",
+    description=(
+        "An offline AI chatbot powered by **Llama-2-7B** (GGMLv3 Q2_K quantized).\n\n"
+        "> Note: The model file `llama-2-7b-chat.ggmlv3.q2_K.bin` must be uploaded to the Space files."
+    ),
+    theme=gr.themes.Soft(
+        primary_hue="blue",
+        secondary_hue="slate",
+    ),
+    examples=[
+        "What is machine learning?",
+        "Write a Python function to reverse a string.",
+        "Explain quantum computing in simple terms.",
+        "What are the planets in the solar system?",
+    ],
+    retry_btn="Retry",
+    undo_btn="Undo",
+    clear_btn="Clear",
+)
+
+if __name__ == "__main__":
+    demo.launch()