fixing APU

Author: soumith

imagenet/main.py

@@ -76,7 +76,7 @@ def forward(self, input):
     train, batch_size=args.batchSize, shuffle=True, num_workers=args.nThreads)
 
 criterion = nn.CrossEntropyLoss().cuda()
-optimizer = torch.optim.SGD(model, learningRate, momentum)
+optimizer = torch.optim.SGD(model.parameters(), learningRate, momentum)
 t = trainer.Trainer(model, criterion, optimizer, train_loader)
 
 t.register_plugin(trainer.plugins.ProgressMonitor())

imagenet/resnet.py

@@ -5,7 +5,7 @@
 def conv3x3(in_planes, out_planes, stride=1):
     "3x3 convolution with padding"
     return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
-                     padding=1, no_bias=True)
+                     padding=1, bias=False)
 
 
 class BasicBlock(nn.Container):
@@ -46,12 +46,12 @@ class Bottleneck(nn.Container):
 
     def __init__(self, inplanes, planes, stride=1, downsample=None):
         super(Bottleneck, self).__init__(
-            conv1=nn.Conv2d(inplanes, planes, kernel_size=1, no_bias=True),
+            conv1=nn.Conv2d(inplanes, planes, kernel_size=1, bias=False),
             bn1=nn.BatchNorm2d(planes),
             conv2=nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
-                            padding=1, no_bias=True),
+                            padding=1, bias=False),
             bn2=nn.BatchNorm2d(planes),
-            conv3=nn.Conv2d(planes, planes * 4, kernel_size=1, no_bias=True),
+            conv3=nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False),
             bn3=nn.BatchNorm2d(planes * 4),
             relu=nn.ReLU(inplace=True),
             downsample=downsample,
@@ -86,7 +86,7 @@ def __init__(self, block, layers):
         self.inplanes = 64
         super(ResNet, self).__init__(
             conv1=nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
-                            no_bias=True),
+                            bias=False),
             bn1=nn.BatchNorm2d(64),
             relu=nn.ReLU(inplace=True),
             maxpool=nn.MaxPool2d(kernel_size=3, stride=2, padding=1),
@@ -110,7 +110,7 @@ def _make_layer(self, block, planes, blocks, stride=1):
         if stride != 1 or self.inplanes != planes * block.expansion:
             downsample = nn.Sequential(
                 nn.Conv2d(self.inplanes, planes * block.expansion,
-                          kernel_size=1, stride=stride, no_bias=True),
+                          kernel_size=1, stride=stride, bias=False),
                 nn.BatchNorm2d(planes * block.expansion),
             )