Update space

Gemma_Model.py

@@ -1,6 +1,7 @@
 from transformers import AutoTokenizer, BitsAndBytesConfig, Gemma3ForCausalLM
 import torch
 import json
+import os
 
 __export__ = ["GemmaLLM"]
 
@@ -16,6 +17,7 @@ class GemmaLLM:
             quantization_config=quantization_config,
             low_cpu_mem_usage=True,
             torch_dtype=torch.float16,
+            token=os.environ.get("GEMMA_TOKEN"),
         ).eval()
 
         self.tokenizer = AutoTokenizer.from_pretrained(model_id)