test only ignore log

Author: felixdittrich92

references/detection/train_pytorch.py

@@ -276,7 +276,9 @@ def main(args):
 
     if args.test_only:
         pbar.write("Running evaluation")
-        val_loss, recall, precision, mean_iou = evaluate(model, val_loader, batch_transforms, val_metric, amp=args.amp)
+        val_loss, recall, precision, mean_iou = evaluate(
+            model, val_loader, batch_transforms, val_metric, amp=args.amp, log=lambda **kwargs: None
+        )
         pbar.write(
             f"Validation loss: {val_loss:.6} (Recall: {recall:.2%} | Precision: {precision:.2%} | "
             f"Mean IoU: {mean_iou:.2%})"