Update app.py

app.py

@@ -4,8 +4,9 @@ os.environ["SPACES_ZERO_GPU"] = "false"
 import gradio as gr
 from transformers import pipeline, AutoTokenizer
 import torch
+import spaces
 
-# Load model and tokenizer
+# Load model and tokenizer globally
 model_name = "DISLab/SummLlama3.2-3B"
 print(f"Loading model: {model_name}")
 
@@ -29,6 +30,7 @@ def format_chat_template(document):
     }]
     return tokenizer.apply_chat_template(row_json, tokenize=False, add_generation_prompt=False)
 
+@spaces.GPU
 def summarize(text):
     """Generate summary using the model"""
     try:
@@ -40,7 +42,7 @@ def summarize(text):
             formatted_input,
             max_new_tokens=512,
             do_sample=True,
-            temperature=0.3,
+            temperature=0.7,
             top_p=0.9,
             return_full_text=False
         )