def test(testloader, model, criterion, epoch, use_cuda): model.eval() accs = np.ones((len(testloader))) * -1000.0 losses = np.ones((len(testloader))) * -1000.0 for batch_idx, (inputs, targets) in enumerate(testloader): if use_cuda: inputs, targets = inputs.cuda(), targets.cuda() inputs, targets = torch.autograd.Variable(inputs, volatile=True), torch.autograd.Variable(targets) outputs = model(inputs) losses[batch_idx] = criterion(outputs, targets).item() accs[batch_idx] = auxil.accuracy(outputs.data, targets.data, topk=(1,))[0].item() return (np.average(losses), np.average(accs))
def train(trainloader, model, criterion, optimizer, epoch, use_cuda):
model.train()
accs = np.ones((len(trainloader))) * -1000.0
losses = np.ones((len(trainloader))) * -1000.0
for batch_idx, (inputs, targets) in enumerate(trainloader):
if use_cuda:
inputs, targets = inputs.cuda(), targets.cuda()
inputs, targets = torch.autograd.Variable(inputs), torch.autograd.Variable(targets)
outputs = model(inputs)
loss = criterion(outputs, targets)
losses[batch_idx] = loss.item()
accs[batch_idx] = auxil.accuracy(outputs.data, targets.data)[0].item()
optimizer.zero_grad()
loss.backward()
optimizer.step()
return (np.average(losses), np.average(accs))