try put pip install out of for div in divs

vi_backbones.py

@@ -6,12 +6,8 @@ import datasets
 import subprocess
 from bs4 import BeautifulSoup
 
-subprocess.call(['pip', 'install', 'torchvision'])
-
 _DBNAME = os.path.basename(__file__).split('.')[0]
-
 _HOMEPAGE = "https://huggingface.co/datasets/george-chou/" + _DBNAME
-
 _URL = 'https://pytorch.org/vision/main/_modules/'
 
 
@@ -103,12 +99,14 @@ class vi_backbones(datasets.GeneratorBasedBuilder):
                 model_page = self._parse_url(url + href)
                 divs = model_page.select('div.viewcode-block')
 
+                subprocess.call(['pip', 'install', 'torchvision'])
+                import torchvision.models as models
+
                 for div in divs:
                     div_id = str(div['id'])
                     if div_id.__contains__('_Weights'):
                         m_ver = div_id.split('_Weight')[0].lower()
 
-                        import torchvision.models as models
                         if not hasattr(models, m_ver):
                             continue