6ac9487579408b1ee452fbb27ef4c781882d085c,examples/mnist_mlp.py,,main,#Any#Any#Any#,14

Before Change

                yh, backprop = model.begin_update(X)
                loss, d_loss = categorical_crossentropy(y, yh)
                backprop(d_loss)
                for name, param, d_param in model.weights:
                    if d_param is not None:
                        optimizer(param, d_param, key=name)
                
            with model.use_params(optimizer.averages):
                dev_acc_avg = model.evaluate(dev_X, dev_Y)
                print("Avg dev.: %.3f" % dev_acc_avg)

After Change

    with model.begin_training(train_X, train_Y) as trainer, optimizer:
        trainer.each_epoch(print_accuracy(dev_X, dev_y))
        for X, y in trainer.iterate(train_X, train_y):
            yh, backprop = model.begin_update(X, drop=trainer.dropout)
            loss, d_loss = categorical_crossentropy(guess, yh)
            backprop(d_loss, optimizer(loss))
    with model.use_params(optimizer.averages):
        print("Avg dev.: %.3f" % model.evaluate(dev_X, dev_Y))
        print("Avg test.: %.3f" % model.evaluate(test_X, test_Y))
        with open("out.pickle", "wb") as file_:

Instances