Update app.py

app.py

@@ -3,11 +3,10 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
 from peft import PeftModel
 import streamlit as st
 
-# Load tokenizer and model (CPU)
+# Load tokenizer and model (on CPU)
 base_model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
-adapter_path = "lora_adapter"  # Your LoRA adapter folder path
+adapter_path = "lora_adapter"
 
-# Force CPU usage
 device = torch.device("cpu")
 
 tokenizer = AutoTokenizer.from_pretrained(base_model_name, use_fast=True)
@@ -15,20 +14,18 @@ base_model = AutoModelForCausalLM.from_pretrained(base_model_name).to(device)
 model = PeftModel.from_pretrained(base_model, adapter_path).to(device)
 
 # Streamlit UI setup
-st.set_page_config(
-    page_title="Python Tutor Chatbot",
-    page_icon="🐍",
-    layout="centered"
-)
-
+st.set_page_config(page_title="Python Tutor Chatbot", page_icon="🐍", layout="centered")
 st.title("🐍 Python Tutor Chatbot")
 st.markdown("Ask me anything about Python programming!")
 
-# Prompt template
+# 🔁 Prompt template with instruction to ignore unrelated queries
 def create_prompt(user_input):
-    return f"""
-You are a helpful and knowledgeable AI Python Tutor. Your job is to answer only Python-related programming questions.
-If the question is unrelated to Python, kindly respond with: "Sorry, I can only answer Python programming questions."
+    return f"""You are a helpful and expert AI Python tutor.
+
+Your job is to only answer questions strictly related to Python programming (syntax, concepts, libraries, frameworks, tools, errors, etc.).
+
+If the question is unrelated to Python, politely respond:
+"Sorry, I can only answer Python programming questions."
 
 ### Instruction:
 {user_input}
@@ -36,25 +33,27 @@ If the question is unrelated to Python, kindly respond with: "Sorry, I can only
 ### Response:
 """
 
-# Chat interface
+# 🔎 User Input
 user_input = st.text_input("Your Python Question:")
 
+# 🔄 Inference
 if user_input:
     with st.spinner("Generating response..."):
         prompt = create_prompt(user_input)
         inputs = tokenizer(prompt, return_tensors="pt").to(device)
 
         with torch.no_grad():
-            output = model.generate(
+            outputs = model.generate(
                 **inputs,
                 max_new_tokens=200,
                 temperature=0.7,
-                do_sample=True,
                 top_p=0.9,
-                top_k=50
+                top_k=50,
+                do_sample=True,
+                pad_token_id=tokenizer.eos_token_id
             )
 
-        response = tokenizer.decode(output[0], skip_special_tokens=True)
+        response = tokenizer.decode(outputs[0], skip_special_tokens=True)
         final_response = response.split("### Response:")[-1].strip()
 
         st.markdown("**Answer:**")