Completed Prototype

Gemma.py

@@ -1,5 +1,6 @@
 from transformers import AutoTokenizer, Gemma3ForCausalLM
 from huggingface_hub import login
+import spaces
 import torch
 import json
 import os
@@ -21,6 +22,7 @@ class GemmaLLM:
             # token=os.environ.get("GEMMA_TOKEN"),
         ).eval()
 
+    @spaces.GPU
     def generate(self, message) -> str:
         inputs = self.tokenizer.apply_chat_template(
             message,

requirements.txt

@@ -4,4 +4,5 @@ newsapi_python==0.2.7
 newspaper3k==0.2.8
 transformers==4.50.0
 lxml_html_clean==0.4.1
-accelerate==1.5.2
+accelerate==1.5.2
+spaces