Update app.py

app.py

@@ -5,11 +5,9 @@ import streamlit as st
 
 st.set_page_config(page_title="TinyLLaMA Python Tutor", layout="centered")
 
-# Title
 st.title("🧠 TinyLLaMA Python Tutor (LoRA)")
 st.write("Ask me any Python programming question:")
 
-# Load base model and LoRA adapter
 @st.cache_resource
 def load_model():
     base_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
@@ -24,16 +22,14 @@ def load_model():
 
 tokenizer, model = load_model()
 
-# Prompt template
 def build_prompt(question):
     return (
-        "You are a helpful AI tutor that only answers Python programming questions. "
-        "If the user asks something unrelated to Python, respond with: "
+        "You are a helpful and concise Python programming tutor. "
+        "If the question is not about Python, respond with: "
         "'Sorry, I can only answer Python-related questions.'\n\n"
         f"Question: {question}\nAnswer:"
     )
 
-# Input box
 question = st.text_input("Your question")
 
 if question:
@@ -43,10 +39,12 @@ if question:
     with st.spinner("Thinking..."):
         outputs = model.generate(
             **inputs,
-            max_new_tokens=300,  # 🔼 Increased from 200 to 300
-            temperature=0.6,      # 🔽 More deterministic
-            top_p=0.9,
-            pad_token_id=tokenizer.eos_token_id
+            max_new_tokens=250,
+            temperature=0.6,
+            top_p=0.85,
+            repetition_penalty=1.2,
+            pad_token_id=tokenizer.eos_token_id,
+            eos_token_id=tokenizer.eos_token_id,
         )
 
     decoded_output = tokenizer.decode(outputs[0], skip_special_tokens=True)