diff --git a/train.py b/train.py
index 4f8e0eb3..f2fde133 100644
--- a/train.py
+++ b/train.py
@@ -372,6 +372,15 @@ def train():
     return results
 
 
+def prebias():
+    # trains output bias layers for 1 epoch and creates new backbone
+    if opt.prebias:
+        train()  # transfer-learn yolo biases for 1 epoch
+        create_backbone(last)  # saved results as backbone.pt
+        opt.weights = wdir + 'backbone.pt'  # assign backbone
+        opt.prebias = False  # disable prebias
+
+
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
     parser.add_argument('--epochs', type=int, default=273)  # 500200 batches at bs 16, 117263 images = 273 epochs
@@ -403,12 +412,6 @@ if __name__ == '__main__':
     device = torch_utils.select_device(opt.device, apex=mixed_precision)
 
     tb_writer = None
-    if opt.prebias:
-        train()  # transfer-learn yolo biases for 1 epoch
-        create_backbone(last)  # saved results as backbone.pt
-        opt.weights = wdir + 'backbone.pt'  # assign backbone
-        opt.prebias = False  # disable prebias
-
     if not opt.evolve:  # Train normally
         try:
             # Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/
@@ -418,6 +421,7 @@ if __name__ == '__main__':
         except:
             pass
 
+        prebias()  # optional
         train()  # train normally
 
     else:  # Evolve hyperparameters (optional)
@@ -455,6 +459,7 @@ if __name__ == '__main__':
                 hyp[k] = np.clip(hyp[k], v[0], v[1])
 
             # Train mutation
+            prebias()
             results = train()
 
             # Write mutation results