loader.set_description(
(f'epoch: {epoch + 1}; loss: {loss.item():.5f}; ' f'acc: {accuracy:.5f}')
)
class PixelTransform:
def __init__(self):
pass
def __call__(self, input):
ar = np.array(input)
return torch.from_numpy(ar).long()
if __name__ == '__main__':
device = 'cuda'
epoch = 10
dataset = datasets.MNIST('.', transform=PixelTransform(), download=True)
loader = DataLoader(dataset, batch_size=32, shuffle=True, num_workers=4)
model = PixelSNAIL([28, 28], 256, 128, 5, 2, 4, 128)
model = model.to(device)
optimizer = optim.Adam(model.parameters(), lr=1e-3)
for i in range(10):
train(i, loader, model, optimizer, device)
torch.save(model.state_dict(), f'checkpoint/mnist_{str(i + 1).zfill(3)}.pt')
(f'epoch: {epoch + 1}; loss: {loss.item():.5f}; '
f'acc: {accuracy:.5f}'))
class PixelTransform:
def __init__(self):
pass
def __call__(self, input):
ar = np.array(input)
return torch.from_numpy(ar).long()
if __name__ == '__main__':
device = 'cuda'
epoch = 10
dataset = datasets.MNIST('.', transform=PixelTransform(), download=True)
loader = DataLoader(dataset, batch_size=32, shuffle=True, num_workers=4)
model = PixelSNAIL([28, 28], 256, 128, 5, 2, 4, 128)
model = model.to(device)
optimizer = optim.Adam(model.parameters(), lr=1e-3)
for i in range(10):
train(i, loader, model, optimizer, device)
torch.save(model.state_dict(),
f'checkpoint/mnist_{str(i + 1).zfill(3)}.pt')