Update app.py

app.py

@@ -7,18 +7,18 @@ import streamlit as st
 base_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
 tokenizer = AutoTokenizer.from_pretrained(base_model)
 
-# Load model in CPU mode
+# Load base model on CPU
 model = AutoModelForCausalLM.from_pretrained(
     base_model,
     torch_dtype=torch.float32,
     device_map="cpu"
 )
 
-# Load the LoRA adapter
+# Load LoRA adapter
 model = PeftModel.from_pretrained(model, "lora_adapter", device_map="cpu")
 model.eval()
 
-# Format prompt with example
+# Prompt template
 def format_prompt(instruction):
     return f"""### SYSTEM:
 You are a helpful and expert Python programming tutor.
@@ -26,12 +26,6 @@ You only answer questions related to Python programming.
 If the question is unrelated to Python, say:
 "Sorry, I can only answer Python-related questions."
 
-### USER:
-What is the difference between a list and a tuple in Python?
-
-### ASSISTANT:
-In Python, both lists and tuples are used to store collections of items, but they have key differences. Lists are mutable (can be changed), whereas tuples are immutable (cannot be changed). Lists use square brackets [], and tuples use parentheses (). Tuples are generally faster and use less memory.
-
 ### USER:
 {instruction}
 
@@ -43,23 +37,26 @@ def chat(instruction):
     prompt = format_prompt(instruction)
     inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
 
-    outputs = model.generate(
-        **inputs,
-        max_new_tokens=256,
-        do_sample=False,
-        temperature=0.0,
-        top_p=1.0,
-        repetition_penalty=1.1
-    )
+    with torch.no_grad():
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=512,
+            do_sample=False,
+            temperature=0.7,
+            top_p=0.9,
+            repetition_penalty=1.1
+        )
 
     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    print("🧠 Raw Model Output:", repr(response))  # Debug log
 
-    # Extract response cleanly
+    # Debugging optional:
+    # print("🧠 Full output:", repr(response))
+
+    # Split and return the assistant's answer
     if "### ASSISTANT:" in response:
         return response.split("### ASSISTANT:")[-1].strip()
     else:
-        return response.strip()  # fallback
+        return response.strip()
 
 # Streamlit UI
 st.set_page_config(page_title="🐍 Python Tutor Chatbot")
@@ -73,7 +70,3 @@ if st.button("Get Answer") and user_input.strip():
         response = chat(user_input)
         st.markdown("**Answer:**")
         st.write(response)
-
-        # Optional debug output
-        # st.write("**Raw model output:**")
-        # st.write(repr(response))