Softmax

用mxnet写softmax, epoch_num没有设置的太大，跑几次看看结果。 因为引入exp函数，在数字太大时会报错，可能不收敛。

代码

import mxnet import d2lzh as d2l from mxnet import autograd batch_size = 256 train_iter, test_iter = d2l.load_data_fashion_mnist((batch_size)) net = mxnet.gluon.nn.Sequential() # 容器 net.add(mxnet.gluon.nn.Flatten()) # 平滑 net.add((mxnet.gluon.nn.Dense(10))) # 全连接层,10个输出 net.initialize(mxnet.init.Normal(sigma=0.01)) # 初始化 loss = mxnet.gluon.loss.SoftmaxCrossEntropyLoss() # 定义损失函数Softmax trainer = mxnet.gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': 0.1}) # 训练器初始化 num_epochs = 5 lr = 0.1 for epoch in range(num_epochs): train_l_sum, train_acc_sum, n = 0.0, 0.0, 0 for X, y in train_iter: with autograd.record(): y_hat = net(X) l = loss(y_hat, y).sum() l.backward() # 求导 trainer.step(batch_size) # 迭代并更新 y = y.astype('float32') train_l_sum += l.asscalar() train_acc_sum += (y_hat.argmax(axis=1) == y).sum().asscalar() n += y.size test_acc_sum, test_n = 0.0, 0 for test_X, test_y in test_iter: test_y = test_y.astype('float32') test_acc_sum += (net(test_X).argmax(axis=1) == test_y).sum().asscalar() test_n += test_y.size test_acc = test_acc_sum / test_n print('epoch {}, loss {:.4f}, train acc {:.3f}, test_acc {:.3f}'.format(epoch + 1, train_l_sum / n, train_acc_sum / n, test_acc))

结果

epoch 1, loss 0.7893, train acc 0.746, test_acc 0.808 epoch 2, loss 0.5745, train acc 0.811, test_acc 0.826 epoch 3, loss 0.5291, train acc 0.824, test_acc 0.827 epoch 4, loss 0.5052, train acc 0.830, test_acc 0.837 epoch 5, loss 0.4896, train acc 0.834, test_acc 0.842