Update app.py

app.py

@@ -64,20 +64,17 @@
 
 
 
-
-
-
 import gradio as gr
 from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
 
 # --- Load Model ---
-MODEL_PATH = "./tinyllama-jobskills-final_update_4"  # Model files path
+MODEL_PATH = "./tinyllama-jobskills-final_update_4"  # Path to your model
 
 tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_PATH,
     trust_remote_code=True,
-    device_map="auto",  # Will use GPU if available
+    device_map="auto",   # Use GPU if available
     low_cpu_mem_usage=True
 )
 
@@ -88,44 +85,54 @@ pipe = pipeline(
     device_map="auto"
 )
 
-# --- Define Chat Function ---
+# --- Chat Function ---
 def chat_fn(message, history):
-    # Convert history to text
+    # Convert history into text for prompt
     history_text = ""
-    for user, bot in history[-5:]:  # Use only last 5 exchanges for speed
-        history_text += f"User: {user}\nAssistant: {bot}\n"
+    for msg in history[-10:]:
+        role = "User" if msg["role"] == "user" else "Assistant"
+        history_text += f"{role}: {msg['content']}\n"
     history_text += f"User: {message}\nAssistant:"
 
-    # Generate response
+    # Generate response from model
     response = pipe(
         history_text,
-        max_new_tokens=64,  # Reduced for CPU
-        do_sample=False,    # Greedy decoding for speed
-        temperature=1.0,
+        max_new_tokens=16,
+        do_sample=False,
+        temperature=0.7,
         top_p=1.0
     )[0]["generated_text"]
 
     # Extract assistant reply
     reply = response.split("Assistant:")[-1].strip()
-    return reply
+
+    # Format reply into bullet points
+    skills = [s.strip() for s in reply.replace(",", "\n").split("\n") if s.strip()]
+    formatted_reply = "\n".join([f"- {s}" for s in skills])
+
+    return formatted_reply
+
 
 # --- Gradio UI ---
 with gr.Blocks() as demo:
-    gr.Markdown("## 🚀 Chat with My Custom Model (CPU-Friendly)")
+    gr.Markdown("## 🚀 Chat with My AI Skills Model")
 
-    chatbot = gr.Chatbot(type="messages")  # updated for future versions
-    msg = gr.Textbox(label="Type your message")
+    chatbot = gr.Chatbot(type="messages")
+    msg = gr.Textbox(label="Type your question here...")
     clear = gr.Button("Clear")
 
     def user_fn(user_message, chat_history):
         bot_message = chat_fn(user_message, chat_history)
-        chat_history.append((user_message, bot_message))
+        chat_history.append({"role": "user", "content": user_message})
+        chat_history.append({"role": "assistant", "content": bot_message})
         return "", chat_history
 
     msg.submit(user_fn, [msg, chatbot], [msg, chatbot])
-    clear.click(lambda: [], None, chatbot, queue=False)  # clears chat
+    clear.click(lambda: [], None, chatbot, queue=False)
 
 # --- Launch ---
 if __name__ == "__main__":
     demo.launch()
 
+
+