动手学深度学习框架怎么选？多框架代码 + 交互式实验环境指南

import tensorflow as tf
from tensorflow.keras import layers

model = tf.keras.Sequential([
    layers.Dense(, activation='relu', input_shape=(,)),
    layers.Dense(, activation='softmax')
])

model.compile(optimizer='adam',
              loss='sparse_categorical_crossentropy',
              metrics=['accuracy'])

model.fit(x_train, y_train, epochs=)

import torch
import torch.nn as nn
import torch.optim as optim

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.fc1 = nn.Linear(, )
        self.fc2 = nn.Linear(, )

    def forward(self, x):
        x = x.view(-, )
        x = torch.relu(self.fc1(x))
        x = self.fc2(x)
        return x

model = Net()
optimizer = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)
criterion = nn.CrossEntropyLoss()

model.train()
for epoch in range():
    running_loss = 0.0
    for inputs, labels in trainloader:
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()
        running_loss += loss.item()
    print(f'Epoch {epoch+}, Loss: {running_loss/len(trainloader)}')

import mxnet as mx
from mxnet import gluon, autograd

net = gluon.nn.Sequential()
with net.name_scope():
    net.add(gluon.nn.Dense(, activation='relu'))
    net.add(gluon.nn.Dense())

net.initialize()
trainer = gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': 0.001})
loss = gluon.loss.SoftmaxCrossEntropyLoss()

for epoch in range():
    running_loss = 0.0
    for data, label in train_data:
        data = data.as_in_context(ctx)
        label = label.as_in_context(ctx)
        with autograd.record():
            output = net(data)
            L = loss(output, label)
        L.backward()
        trainer.step(data.shape[])
        running_loss += L.mean().asscalar()
    print(f'Epoch {epoch+}, Loss: {running_loss/len(train_data)}')

pip install jupyter notebook