神经网络分类任务python入门

目录

Mnist分类任务

读取Mnist数据集

转换成tensor才能参与后续建模训练

torch.nn.functional

创建一个model来更简化代码

使用TensorDataset和DataLoader来简化

整个过程

---

Mnist分类任务

- 网络基本构建与训练方法，常用函数解析

- torch.nn.functional模块

- nn.Module模块

读取Mnist数据集

- 会自动进行下载

from pathlib import Path
import requests

DATA_PATH = Path("data")
PATH = DATA_PATH / "mnist"

PATH.mkdir(parents=True, exist_ok=True)

URL = "http://deeplearning.net/data/mnist/"
FILENAME = "mnist.pkl.gz"

if not (PATH / FILENAME).exists():
        content = requests.get(URL + FILENAME).content
        (PATH / FILENAME).open("wb").write(content)

 

import pickle
import gzip

with gzip.open((PATH / FILENAME).as_posix(), "rb") as f:
        ((x_train, y_train), (x_valid, y_valid), _) = pickle.load(f, encoding="latin-1")

from matplotlib import pyplot
import numpy as np

pyplot.imshow(x_train[0].reshape((28, 28)), cmap="gray")
print(x_train.shape)

转换成tensor才能参与后续建模训练

import torch

x_train, y_train, x_valid, y_valid = map(
    torch.tensor, (x_train, y_train, x_valid, y_valid)
)
n, c = x_train.shape
x_train, x_train.shape, y_train.min(), y_train.max()
print(x_train, y_train)
print(x_train.shape)
print(y_train.min(), y_train.max())

tensor([[0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], ..., [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.], [0., 0., 0., ..., 0., 0., 0.]]) tensor([5, 0, 4, ..., 8, 4, 8]) torch.Size([50000, 784]) tensor(0) tensor(9)

torch.nn.functional

很多层和函数在这里都会见到

torch.nn.functional中有很多功能，后续会常用的。那什么时候使用nn.Module，什么时候使用nn.functional呢？一般情况下，如果模型有可学习的参数，最好用nn.Module，其他情况nn.functional相对更简单一些

import torch.nn.functional as F

loss_func = F.cross_entropy

def model(xb):
    return xb.mm(weights) + bias

bs = 64
xb = x_train[0:bs]  # a mini-batch from x
yb = y_train[0:bs]
weights = torch.randn([784, 10], dtype = torch.float,  requires_grad = True) 
bs = 64
bias = torch.zeros(10, requires_grad=True)

print(loss_func(model(xb), yb))

创建一个model来更简化代码

- 必须继承nn.Module且在其构造函数中需调用nn.Module的构造函数

- 无需写反向传播函数，nn.Module能够利用autograd自动实现反向传播

- Module中的可学习参数可以通过named_parameters()或者parameters()返回迭代器

from torch import nn

class Mnist_NN(nn.Module):
    def __init__(self):
        super().__init__()
        self.hidden1 = nn.Linear(784, 128)
        self.hidden2 = nn.Linear(128, 256)
        self.out  = nn.Linear(256, 10)

    def forward(self, x):
        x = F.relu(self.hidden1(x))
        x = F.relu(self.hidden2(x))
        x = self.out(x)
        return x
        

net = Mnist_NN()

print(net)

Mnist_NN( (hidden1): Linear(in_features=784, out_features=128, bias=True) (hidden2): Linear(in_features=128, out_features=256, bias=True) (out): Linear(in_features=256, out_features=10, bias=True) )

打印定义好名字里的权重和偏置项

for name, parameter in net.named_parameters():

    print(name, parameter,parameter.size())

Output exceeds the size limit. Open the full output data in a text editor

hidden1.weight Parameter containing: tensor([[ 0.0018, 0.0218, 0.0036, ..., -0.0286, -0.0166, 0.0089], [-0.0349, 0.0268, 0.0328, ..., 0.0263, 0.0200, -0.0137], [ 0.0061, 0.0060, -0.0351, ..., 0.0130, -0.0085, 0.0073], ..., [-0.0231, 0.0195, -0.0205, ..., -0.0207, -0.0103, -0.0223], [-0.0299, 0.0305, 0.0098, ..., 0.0184, -0.0247, -0.0207], [-0.0306, -0.0252, -0.0341, ..., 0.0136, -0.0285, 0.0057]], requires_grad=True) torch.Size([128, 784]) hidden1.bias Parameter containing: tensor([ 0.0072, -0.0269, -0.0320, -0.0162, 0.0102, 0.0189, -0.0118, -0.0063, -0.0277, 0.0349, 0.0267, -0.0035, 0.0127, -0.0152, -0.0070, 0.0228, -0.0029, 0.0049, 0.0072, 0.0002, -0.0356, 0.0097, -0.0003, -0.0223, -0.0028, -0.0120, -0.0060, -0.0063, 0.0237, 0.0142, 0.0044, -0.0005, 0.0349, -0.0132, 0.0138, -0.0295, -0.0299, 0.0074, 0.0231, 0.0292, -0.0178, 0.0046, 0.0043, -0.0195, 0.0175, -0.0069, 0.0228, 0.0169, 0.0339, 0.0245, -0.0326, -0.0260, -0.0029, 0.0028, 0.0322, -0.0209, -0.0287, 0.0195, 0.0188, 0.0261, 0.0148, -0.0195, -0.0094, -0.0294, -0.0209, -0.0142, 0.0131, 0.0273, 0.0017, 0.0219, 0.0187, 0.0161, 0.0203, 0.0332, 0.0225, 0.0154, 0.0169, -0.0346, -0.0114, 0.0277, 0.0292, -0.0164, 0.0001, -0.0299, -0.0076, -0.0128, -0.0076, -0.0080, -0.0209, -0.0194, -0.0143, 0.0292, -0.0316, -0.0188, -0.0052, 0.0013, -0.0247, 0.0352, -0.0253, -0.0306, 0.0035, -0.0253, 0.0167, -0.0260, -0.0179, -0.0342, 0.0033, -0.0287, -0.0272, 0.0238, 0.0323, 0.0108, 0.0097, 0.0219, 0.0111, 0.0208, -0.0279, 0.0324, -0.0325, -0.0166,

...

requires_grad=True) torch.Size([10, 256]) out.bias Parameter containing: tensor([-0.0536, 0.0007, 0.0227, -0.0072, -0.0168, -0.0125, -0.0207, -0.0558, 0.0579, -0.0439], requires_grad=True) torch.Size([10])

使用TensorDataset和DataLoader来简化

from torch.utils.data import TensorDataset
from torch.utils.data import DataLoader

train_ds = TensorDataset(x_train, y_train)
train_dl = DataLoader(train_ds, batch_size=bs, shuffle=True)

valid_ds = TensorDataset(x_valid, y_valid)
valid_dl = DataLoader(valid_ds, batch_size=bs * 2)

def get_data(train_ds, valid_ds, bs):
    return (
        DataLoader(train_ds, batch_size=bs, shuffle=True),
        DataLoader(valid_ds, batch_size=bs * 2),
    )

- 一般在训练模型时加上model.train()，这样会正常使用Batch Normalization和 Dropout

- 测试的时候一般选择model.eval()，这样就不会使用Batch Normalization和 Dropout

import numpy as np

def fit(steps, model, loss_func, opt, train_dl, valid_dl):
    for step in range(steps):
        model.train()
        for xb, yb in train_dl:
            loss_batch(model, loss_func, xb, yb, opt)

        model.eval()
        with torch.no_grad():
            losses, nums = zip(
                *[loss_batch(model, loss_func, xb, yb) for xb, yb in valid_dl]
            )
        val_loss = np.sum(np.multiply(losses, nums)) / np.sum(nums)
        print('当前step:'+str(step), '验证集损失：'+str(val_loss))

from torch import optim
def get_model():
    model = Mnist_NN()
    return model, optim.SGD(model.parameters(), lr=0.001)

def loss_batch(model, loss_func, xb, yb, opt=None):
    loss = loss_func(model(xb), yb)

    if opt is not None:
        loss.backward()
        opt.step()
        opt.zero_grad()

    return loss.item(), len(xb)

整个过程

train_dl, valid_dl = get_data(train_ds, valid_ds, bs)

model, opt = get_model()

fit(25, model, loss_func, opt, train_dl, valid_dl)