Updated for CPU usage3

app.py

@@ -1,14 +1,12 @@
 import streamlit as st
-from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline 
+from transformers import AutoTokenizer, AutoModelForCausalLM
 
 # Use a smaller model 
 model_name = "microsoft/DialoGPT-small"  
 
 # Use the model's default GPT2 tokenizer
 tokenizer = AutoTokenizer.from_pretrained(model_name)
-
 model = AutoModelForCausalLM.from_pretrained(model_name)
-chat_pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer, device=-1, max_length=1000)
 
 # Streamlit UI
 st.title("Buddy Christ Chatbot 🌟")
@@ -21,7 +19,7 @@ if user_input:
     """
     # Tokenize the input with truncation
     inputs = tokenizer(buddy_christ_prefix + user_input, return_tensors="pt", truncation=True, max_length=1000)
-    # Generate the response
-    response = model.generate(**inputs)
+    # Generate the response with temperature and top-k sampling
+    response = model.generate(**inputs, max_length=1000, temperature=0.8, top_k=50)
     response_text = tokenizer.decode(response[0], skip_special_tokens=True)
     st.write("Buddy Christ:", response_text[len(buddy_christ_prefix):])