diff --git a/models.py b/models.py
index eb2fdcea..119aaf4b 100755
--- a/models.py
+++ b/models.py
@@ -96,7 +96,7 @@ class YOLOLayer(nn.Module):
     def __init__(self, anchors, nc, img_size, yolo_index):
         super(YOLOLayer, self).__init__()
 
-        self.anchors = torch.Tensor(anchors)
+        self.anchors = torch.from_numpy(anchors)
         self.na = len(anchors)  # number of anchors (3)
         self.nc = nc  # number of classes (80)
         self.nx = 0  # initialize number of x gridpoints
diff --git a/train.py b/train.py
index 0166dccb..c0103aa2 100644
--- a/train.py
+++ b/train.py
@@ -200,8 +200,7 @@ def train(cfg,
     # Start training
     model.nc = nc  # attach number of classes to model
     model.hyp = hyp  # attach hyperparameters to model
-    if dataset.image_weights:
-        model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device)  # attach class weights
+    model.class_weights = labels_to_class_weights(dataset.labels, nc).to(device)  # attach class weights
     model_info(model, report='summary')  # 'full' or 'summary'
     nb = len(dataloader)
     maps = np.zeros(nc)  # mAP per class
diff --git a/utils/utils.py b/utils/utils.py
index 0a0ed43c..497dab38 100755
--- a/utils/utils.py
+++ b/utils/utils.py
@@ -62,7 +62,7 @@ def labels_to_class_weights(labels, nc=80):
     weights[weights == 0] = 1  # replace empty bins with 1
     weights = 1 / weights  # number of targets per class
     weights /= weights.sum()  # normalize
-    return torch.Tensor(weights)
+    return torch.from_numpy(weights)
 
 
 def labels_to_image_weights(labels, nc=80, class_weights=np.ones(80)):