add reduce mean for loss

cnn_e2e/of_cnn_train_val.py

@@ -51,7 +51,7 @@ def TrainNet():
 
     logits = model_dict[args.model](images)
     loss = flow.nn.sparse_softmax_cross_entropy_with_logits(labels, logits, name="softmax_loss")
-    #loss = flow.math.reduce_mean(loss)
+    loss = flow.math.reduce_mean(loss)
     flow.losses.add_loss(loss)
     predictions = flow.nn.softmax(logits)
     outputs = {"loss": loss, "predictions":predictions, "labels": labels}
@@ -84,8 +84,8 @@ def main():
     snapshot = Snapshot(args.model_save_dir, args.model_load_dir)
 
     for epoch in range(args.num_epochs):
-        metric = Metric(desc='train', calculate_batches=args.loss_print_every_n_iter, 
-                        summary=summary, save_summary_steps=epoch_size, 
+        metric = Metric(desc='train', calculate_batches=args.loss_print_every_n_iter,
+                        summary=summary, save_summary_steps=epoch_size,
                         batch_size=train_batch_size, loss_key='loss')
         for i in range(epoch_size):
             TrainNet().async_get(metric.metric_cb(epoch, i))
@@ -93,7 +93,7 @@ def main():
         #        break
         #break
         if args.val_data_dir:
-            metric = Metric(desc='validataion', calculate_batches=num_val_steps, summary=summary, 
+            metric = Metric(desc='validataion', calculate_batches=num_val_steps, summary=summary,
                             save_summary_steps=num_val_steps, batch_size=val_batch_size)
             for i in range(num_val_steps):
                 InferenceNet().async_get(metric.metric_cb(epoch, i))