Commit 236184f

authored

Fixing invalid index of a 0-dim tensor.

Fixing invalid index of a 0-dim tensor. This will be an error in PyTorch 0.5. Use tensor.item() to convert a 0-dim tensor to a Python number

1 parent 29f5b4f commit 236184fCopy full SHA for 236184f

File tree

-4

lines changed

-4

lines changed

Lines changed: 4 additions & 4 deletions

Original file line number	Diff line number	Diff line change
`@@ -507,10 +507,10 @@`
`507`	`507`	`" loss.backward()\n",`
`508`	`508`	`" optimizer.step()\n",`
`509`	`509`	`" # 记录误差\n",`
`510`		`- " train_loss += loss.data[0]\n",`
	`510`	`+ " train_loss += loss.item()\n",`
`511`	`511`	`" # 计算分类的准确率\n",`
`512`	`512`	`" _, pred = out.max(1)\n",`
`513`		`- " num_correct = (pred == label).sum().data[0]\n",`
	`513`	`+ " num_correct = (pred == label).sum().item()\n",`
`514`	`514`	`" acc = num_correct / im.shape[0]\n",`
`515`	`515`	`" train_acc += acc\n",`
`516`	`516`	`" \n",`
`@@ -526,10 +526,10 @@`
`526`	`526`	`" out = net(im)\n",`
`527`	`527`	`" loss = criterion(out, label)\n",`
`528`	`528`	`" # 记录误差\n",`
`529`		`- " eval_loss += loss.data[0]\n",`
	`529`	`+ " eval_loss += loss.item()\n",`
`530`	`530`	`" # 记录准确率\n",`
`531`	`531`	`" _, pred = out.max(1)\n",`
`532`		`- " num_correct = (pred == label).sum().data[0]\n",`
	`532`	`+ " num_correct = (pred == label).sum().item()\n",`
`533`	`533`	`" acc = num_correct / im.shape[0]\n",`
`534`	`534`	`" eval_acc += acc\n",`
`535`	`535`	`" \n",`

Comments

(0)