Update space

Gemma_Model.py

@@ -1,4 +1,8 @@
-from transformers import AutoTokenizer, BitsAndBytesConfig, Gemma3ForCausalLM
+from transformers import (
+    AutoTokenizer,
+    # BitsAndBytesConfig,
+    Gemma3ForCausalLM,
+    )
 import torch
 import json
 import os
@@ -9,14 +13,14 @@ class GemmaLLM:
     
     def __init__(self):
         model_id = "google/gemma-3-1b-it"
-        quantization_config = BitsAndBytesConfig(load_in_8bit=True)
+        # quantization_config = BitsAndBytesConfig(load_in_8bit=True)
         
         self.model = Gemma3ForCausalLM.from_pretrained(
             model_id,
             device_map="cpu",
-            quantization_config=quantization_config,
-            low_cpu_mem_usage=True,
-            torch_dtype=torch.float16,
+            # quantization_config=quantization_config,
+            # low_cpu_mem_usage=True,
+            # torch_dtype=torch.float16,
             token=os.environ.get("GEMMA_TOKEN"),
         ).eval()
 

requirements.txt

@@ -5,5 +5,4 @@ newspaper3k==0.2.8
 torch==2.6.0
 transformers==4.50.0
 lxml_html_clean==0.4.1
-accelerate==1.5.2
-bitsandbytes==0.45.3
+accelerate==1.5.2