Update logic to rescale L2 loss in distribution strategy (#6601)

official/resnet/resnet_run_loop.py

@@ -419,16 +419,14 @@ def resnet_model_fn(features, labels, mode, model_class,
     return 'batch_normalization' not in name
   loss_filter_fn = loss_filter_fn or exclude_batch_norm
 
-  # Add weight decay to the loss. We need to scale the regularization loss
-  # manually as losses other than in tf.losses and tf.keras.losses don't scale
-  # automatically.
+  # Add weight decay to the loss.
   l2_loss = weight_decay * tf.add_n(
       # loss is computed using fp32 for numerical stability.
       [
           tf.nn.l2_loss(tf.cast(v, tf.float32))
           for v in tf.compat.v1.trainable_variables()
           if loss_filter_fn(v.name)
-      ]) / tf.distribute.get_strategy().num_replicas_in_sync
+      ])
   tf.compat.v1.summary.scalar('l2_loss', l2_loss)
   loss = cross_entropy + l2_loss