Fix validation loss logging (open-mmlab#1494)

Author: mfernezir

mmseg/models/segmentors/base.py

@@ -155,9 +155,14 @@ def val_step(self, data_batch, optimizer=None, **kwargs):
         losses = self(**data_batch)
         loss, log_vars = self._parse_losses(losses)
 
+        log_vars_ = dict()
+        for loss_name, loss_value in log_vars.items():
+            k = loss_name + '_val'
+            log_vars_[k] = loss_value
+
         outputs = dict(
             loss=loss,
-            log_vars=log_vars,
+            log_vars=log_vars_,
             num_samples=len(data_batch['img_metas']))
 
         return outputs