Update examples to new API (#205)

examples/efficientnet.py

@@ -38,7 +38,7 @@ def infer(model, img):
 
   # run the net
   if GPU:
-    out = model.forward(Tensor(img).cuda()).cpu()
+    out = model.forward(Tensor(img).gpu()).cpu()
   else:
     out = model.forward(Tensor(img))
 
@@ -55,7 +55,7 @@ if __name__ == "__main__":
   model = EfficientNet(int(os.getenv("NUM", "0")))
   model.load_weights_from_torch()
   if GPU:
-    [x.cuda_() for x in get_parameters(model)]
+    [x.gpu_() for x in get_parameters(model)]
 
   # category labels
   import ast

examples/mnist_gan.py

@@ -65,7 +65,7 @@ if __name__ == "__main__":
   ds_noise = Tensor(np.random.randn(64,128).astype(np.float32), gpu=GPU, requires_grad=False)
   n_steps = int(train_data_size/batch_size)
   if GPU:
-    [x.cuda_() for x in generator_params+discriminator_params]
+    [x.gpu_() for x in generator_params+discriminator_params]
   # optimizers
   optim_g = optim.Adam(generator_params,lr=0.0002, b1=0.5) # 0.0002 for equilibrium!
   optim_d = optim.Adam(discriminator_params,lr=0.0002, b1=0.5)

examples/serious_mnist.py

@@ -85,7 +85,7 @@ class BigConvNet:
         try:
           par.cpu().data[:] = np.load(f)
           if GPU:
-            par.cuda()
+            par.gpu()
         except:
           print('Could not load parameter')
 
@@ -126,7 +126,7 @@ if __name__ == "__main__":
 
   if GPU:
     params = get_parameters(model)
-    [x.cuda_() for x in params]
+    [x.gpu_() for x in params]
 
   for lr, epochs in zip(lrs, epochss):
     optimizer = optim.Adam(model.parameters(), lr=lr)