Update 504_batch_normalization.ipynb

Found 2 Errors that does not activate in Colab

Author: changmeen

tutorial-contents-notebooks/504_batch_normalization.ipynb

@@ -96,7 +96,7 @@
     "test_x = Variable(torch.from_numpy(test_x).float(), volatile=True)  # not for computing gradients\n",
     "test_y = Variable(torch.from_numpy(test_y).float(), volatile=True)\n",
     "\n",
-    "train_dataset = Data.TensorDataset(data_tensor=train_x, target_tensor=train_y)\n",
+    "train_dataset = Data.TensorDataset(train_x, train_y)\n",
     "train_loader = Data.DataLoader(dataset=train_dataset, batch_size=BATCH_SIZE, shuffle=True, num_workers=2,)\n",
     "\n",
     "# show data\n",
@@ -274,7 +274,7 @@
     "    for net, l in zip(nets, losses):\n",
     "        net.eval()              # set eval mode to fix moving_mean and moving_var\n",
     "        pred, layer_input, pre_act = net(test_x)\n",
-    "        l.append(loss_func(pred, test_y).data[0])\n",
+    "        l.append(loss_func(pred, test_y).data)\n",
     "        layer_inputs.append(layer_input)\n",
     "        pre_acts.append(pre_act)\n",
     "        net.train()             # free moving_mean and moving_var\n",