diff --git a/Run.py b/Run.py
index 793eafb..607c2c9 100644
--- a/Run.py
+++ b/Run.py
@@ -55,6 +55,9 @@ def str2bool(v):
 
 parser.add_argument("--hide-debug", action="store_true",
                     help="Supress the printing of each 3d location")
+parser.add_argument("--run-on-cpu", default=False, action="store_true",
+                    help="Runs the code on CPU")
+
 
 
 def plot_regressed_3d_bbox(img, cam_to_img, box_2d, dimensions, alpha, theta_ray, img_2d=None):
@@ -85,8 +88,12 @@ def main():
         print('Using previous model %s'%model_lst[-1])
         my_vgg = vgg.vgg19_bn(pretrained=True)
         # TODO: load bins from file or something
-        model = Model.Model(features=my_vgg.features, bins=2).cuda()
-        checkpoint = torch.load(weights_path + '/%s'%model_lst[-1])
+        if FLAGS.run_on_cpu == True :
+            model = Model.Model(features=my_vgg.features, bins=2)#.cuda()
+            checkpoint = torch.load(weights_path + '/%s'%model_lst[-1], map_location=torch.device('cpu'))
+        else:
+            model = Model.Model(features=my_vgg.features, bins=2).cuda
+            checkpoint = torch.load(weights_path + '/%s'%model_lst[-1])
         model.load_state_dict(checkpoint['model_state_dict'])
         model.eval()
 
@@ -154,7 +161,11 @@ def main():
             box_2d = detection.box_2d
             detected_class = detection.detected_class
 
-            input_tensor = torch.zeros([1,3,224,224]).cuda()
+            if FLAGS.run_on_cpu == True :
+
+                input_tensor = torch.zeros([1,3,224,224])  #.cuda()
+            else:
+                input_tensor = torch.zeros([1,3,224,224]).cuda()
             input_tensor[0,:,:,:] = input_img
 
             [orient, conf, dim] = model(input_tensor)