loss_function = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(params=model_1.parameters(),
 lr=0.01)