【Task6(2天)】PyTorch理解更多神經網絡優化方法
1.瞭解不同優化器
2.書寫優化器代碼
3.Momentum
4.二維優化，隨機梯度下降法進行優化實現
5.Ada自適應梯度調節法
6.RMSProp
7.Adam
8.PyTorch種優化器選擇
2.優化器代碼爲：
import torch
import torch.utils.data as Data
import torch.nn.functional as F
import matplotlib.pyplot as plt
torch.manual_seed(1) # reproducible
LR = 0.01
BATCH_SIZE = 32
EPOCH = 12
# fake dataset
x = torch.unsqueeze(torch.linspace(-1, 1, 1000), dim=1)
y = x.pow(2) + 0.1*torch.normal(torch.zeros(*x.size()))
# plot dataset
plt.scatter(x.numpy(), y.numpy())
plt.show()
# 使用上節內容提到的 data loader
torch_dataset = Data.TensorDataset(x, y)
loader = Data.DataLoader(dataset=torch_dataset, batch_size=BATCH_SIZE, shuffle=True, num_workers=2,)
# 默認的 network 形式
class Net(torch.nn.Module):
def __init__(self):
super(Net, self).__init__()
self.hidden = torch.nn.Linear(1, 20) # hidden layer
self.predict = torch.nn.Linear(20, 1) # output layer
def forward(self, x):
x = F.relu(self.hidden(x)) # activation function for hidden layer
x = self.predict(x) # linear output
return x
# 爲每個優化器創建一個 net
net_SGD = Net()
net_Momentum = Net()
net_RMSprop = Net()
net_Adam = Net()
nets = [net_SGD, net_Momentum, net_RMSprop, net_Adam]
# different optimizers
opt_SGD = torch.optim.SGD(net_SGD.parameters(), lr=LR)
opt_Momentum = torch.optim.SGD(net_Momentum.parameters(), lr=LR, momentum=0.8)
opt_RMSprop = torch.optim.RMSprop(net_RMSprop.parameters(), lr=LR, alpha=0.9)
opt_Adam = torch.optim.Adam(net_Adam.parameters(), lr=LR, betas=(0.9, 0.99))
optimizers = [opt_SGD, opt_Momentum, opt_RMSprop, opt_Adam]
loss_func = torch.nn.MSELoss()
losses_his = [[], [], [], []] # 記錄 training 時不同神經網絡的 loss
for epoch in range(EPOCH):
print('Epoch: ', epoch)
for step, (b_x, b_y) in enumerate(loader):
# 對每個優化器, 優化屬於他的神經網絡
for net, opt, l_his in zip(nets, optimizers, losses_his):
output = net(b_x) # get output for every net
loss = loss_func(output, b_y) # compute loss for every net
opt.zero_grad() # clear gradients for next train
loss.backward() # backpropagation, compute gradients
opt.step() # apply gradients
l_his.append(loss.data.numpy()) # loss recoder
3.Momentum動量法：
#定義動量梯度下降函數
def sgd_momentum(parameters, vs, lr, gamma):
for param, v in zip(parameters, vs):
v[:] = gamma * v + lr * param.grad.data
param.data = param.data - v
import numpy as np
import torch
from torchvision.datasets import MNIST # 導入 pytorch 內置的 mnist 數據
from torch.utils.data import DataLoader
from torch import nn
from torch.autograd import Variable
import time
import matplotlib.pyplot as plt
%matplotlib inline
def data_tf(x):
x = np.array(x, dtype='float32') / 255
x = (x - 0.5) / 0.5 # 標準化，這個技巧之後會講到
x = x.reshape((-1,)) # 拉平
x = torch.from_numpy(x)
return x
train_set = MNIST('./data', train=True, transform=data_tf, download=True) # 載入數據集，申明定義的數據變換
test_set = MNIST('./data', train=False, transform=data_tf, download=True)
# 定義 loss 函數
criterion = nn.CrossEntropyLoss()
train_data = DataLoader(train_set, batch_size=64, shuffle=True)
# 使用 Sequential 定義 3 層神經網絡
net = nn.Sequential(
nn.Linear(784, 200),
nn.ReLU(),
nn.Linear(200, 10),
)
# 將速度初始化爲和參數形狀相同的零張量
vs = []
for param in net.parameters():
vs.append(torch.zeros_like(param.data))
# 開始訓練
losses = []
start = time.time() # 記時開始
for e in range(5):
train_loss = 0
for im, label in train_data:
im = Variable(im)
label = Variable(label)
# 前向傳播
out = net(im)
loss = criterion(out, label)
# 反向傳播
net.zero_grad()
loss.backward()
sgd_momentum(net.parameters(), vs, 1e-2, 0.9) # 使用的動量參數爲 0.9，學習率 0.01
# 記錄誤差
train_loss += loss.data[0]
losses.append(loss.data[0])
print('epoch: {}, Train Loss: {:.6f}'
.format(e, train_loss / len(train_data)))
end = time.time() # 計時結束
print('使用時間: {:.5f} s'.format(end - start))