1 softmax 多分类问题
多分类的交叉熵损失函数
函数知识
1.给类别编码 pd.factorize(data.Species) # 对species编码 # (array([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, 0, 0, 0, 0, 0, 0, 0, 0, # 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, # 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, # 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, # 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, # 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2], dtype=int64), # Index(['setosa', 'versicolor', 'virginica'], dtype='object'))
2
from sklearn.model_selection import train_test_split
train_x, test_x, train_y, test_y = train_test_split(X, Y) # 分训练和测试集
绘图 plt.plot(range(1, epochs+1), train_loss, label='train_loss') plt.plot(range(1, epochs+1), test_loss, label='test_loss') plt.legend() # 添加图例 plt.show()
torch.argmax()函数
Pytorch中torch.argmax()函数解析_cv_lhp的博客-CSDN博客_argmax pytorch
# Softmax多分类
import torch
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import torch.nn.functional as F
from sklearn.model_selection import train_test_split
from torch.utils.data import TensorDataset
from torch.utils.data import DataLoader
from torch import nn
data = pd.read_csv('daatset/iris.csv')
data.Species.unique() # ['setosa' 'versicolor' 'virginica']
pd.factorize(data.Species) # 对species编码
# (array([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, 0, 0, 0, 0, 0, 0, 0, 0,
# 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
# 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
# 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
# 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
# 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2], dtype=int64),
# Index(['setosa', 'versicolor', 'virginica'], dtype='object'))
data['Species'] = pd.factorize(data.Species)[0] # 把类别用对应编码替换
# print(data)
X = data.iloc[:, 1:-1].values # 第一个‘：’不包括第一列，‘1：-1‘表示从第二列开始，最后一个不要
Y = data.Species.values
train_x, test_x, train_y, test_y = train_test_split(X, Y) # 分训练和测试集
train_x = torch.from_numpy(train_x).type(torch.float32)
test_x = torch.from_numpy(test_x).type(torch.float32)
train_y = torch.from_numpy(train_y).type(torch.int64)
test_y = torch.from_numpy(test_y).type(torch.int64)
batch = 8
train_ds = TensorDataset(train_x, train_y)
train_dl = DataLoader(train_ds, batch_size=batch, shuffle=True)
test_ds = TensorDataset(test_x, test_y)
test_dl = DataLoader(test_ds, batch_size=batch)
class Model(nn.Module):
# 定义一个类，从nn.Module中继承
def __init__(self): # 初始化
super().__init__() # 继承父类中所有属性
self.liner_1 = nn.Linear(4, 32) # 输出20层 输出到隐藏层，创建64个隐藏层
self.liner_2 = nn.Linear(32, 32)
self.liner_3 = nn.Linear(32, 3) # 输出层1
def forward(self, input):
# 在forward中使用上面的层，参数为self，对input进行处理
x = F.relu(self.liner_1(input)) # 第一层调用,直接用F调用relu
x = F.relu(self.liner_2(x))
x = self.liner_3(x) #
return x
model = Model()
loss_fn = nn.CrossEntropyLoss()
optim = torch.optim.Adam(model.parameters(), lr=0.0001)
#如何将预测结果转换为真实对应数据
# input_batch, label_batch = next(iter(train_dl))
# input_batch.shape
# label_batch.shape # (torch.Size([8, 4]), torch.Size([8]))
# y_pred = model(input_batch)
# y_pred.shape # torch.Size([8, 3])
# y_pred
# # tensor([[ 0.2699, -0.0188, -0.0153],
# # [ 0.2366, -0.0078, -0.0199],
# # [ 0.0645, 0.0852, 0.0494],
# # [ 0.0868, 0.0561, 0.0322],
# # [ 0.2135, 0.0178, -0.0141],
# # [ 0.0795, 0.0476, 0.0091],
# # [ 0.3266, -0.0554, 0.0235],
# # [ 0.0940, 0.0248, -0.0051]], grad_fn=<AddmmBackward>)
# torch.argmax(y_pred, dim=1) # tensor([0, 0, 1, 0, 0, 0, 0, 0])
# print(a)
# 创建训练函数
def accuracy(y_pred,y_true):
y_pred = torch.argmax(y_pred, dim=1) # 转换为实际的预测结果
acc = (y_pred == y_true).float().mean() # 计算平均正确率
return acc
train_loss = []
train_acc = []
test_loss = []
test_acc = []
epochs = 20
for epoch in range(epochs):
for x, y in train_dl:
y_pred = model(x)
loss = loss_fn(y_pred.to(torch.float32), y.to(torch.int64))
optim.zero_grad()
loss.backward()
optim.step() # 优化
with torch.no_grad():
epoch_accuracy = accuracy(model(train_x),train_y)
epoch_loss = loss_fn(model(train_x).to(torch.float32), train_y.to(torch.int64)).data
epoch_test_accuracy = accuracy(model(test_x), test_y)
epoch_test_loss = loss_fn(model(test_x).to(torch.float32), test_y.to(torch.int64)).data
print('epoch', epoch, 'loss: ', round(epoch_loss.item(), 3), 'accuracy:', round(epoch_accuracy.item(),3),
'test_loss: ', round(epoch_test_loss.item(),3), 'test_accuracy:', round(epoch_test_accuracy.item(),3)
)
train_loss.append(epoch_loss)
train_acc.append(epoch_accuracy)
test_loss.append(epoch_test_loss)
test_acc.append(epoch_test_accuracy)
plt.plot(range(1, epochs+1), train_loss, label='train_loss')
plt.plot(range(1, epochs+1), test_loss, label='test_loss')
plt.legend() # 添加图例
plt.show()
编写fit函数
输入模型、输入数据，对数据输入在模型上训练，并返回loss和acc
import torch
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import torch.nn.functional as F
from sklearn.model_selection import train_test_split
from torch.utils.data import TensorDataset
from torch.utils.data import DataLoader
from torch import nn
def fit(epoch, model, trainloader, testloater):
correct = 0 # correct预测成功数
total = 0 # total训练总样本数
running_loss = 0
for x, y in trainloader:
y_pred = model(x)
loss = loss_fn(y_pred.to(torch.float32), y.to(torch.int64))
optim.zero_grad()
loss.backward()
optim.step() # 优化
with torch.no_grad():
y_pred = torch.argmax(y_pred, dim=1) # argmax dim=1返回y_pred每一行最大数值对应的列号
correct += (y_pred == y).sum().item() # 累加正确样本个数
total += y.size(0) # 总样本数
running_loss += loss.item() # 表示每一个batch的loss
epoch_acc = correct / total
epoch_loss = running_loss / len(trainloader.dataset) # 平均loss trainloader.dataset 返回对应的ds
test_correct = 0 # correct预测成功数
test_total = 0 # total训练总样本数
test_running_loss = 0
with torch.no_grad():
for x, y in testloater:
y_pred = model(x)
loss = loss_fn(y_pred.to(torch.float32), y.to(torch.int64))
y_pred = torch.argmax(y_pred, dim=1) # argmax dim=1返回y_pred每一行最大数值对应的列号
test_correct += (y_pred == y).sum().item() # 累加正确样本个数
test_total += y.size(0) # 总样本数
test_running_loss += loss.item() # 表示每一个batch的loss
epoch_test_loss = test_running_loss / len(testloater.dataset)
epoch_test_acc = test_correct / test_total
print('epoch', epoch,
'loss: ', round(epoch_loss, 3),
'accuracy:', round(epoch_acc, 3),
'test_loss: ', round(epoch_test_loss, 3),
'test_accuracy:', round(epoch_test_acc, 3)
)
return epoch_loss, epoch_acc, epoch_test_loss, epoch_test_acc
data = pd.read_csv('daatset/iris.csv')
data.Species.unique() # ['setosa' 'versicolor' 'virginica']
pd.factorize(data.Species) # 对species编码
data['Species'] = pd.factorize(data.Species)[0] # 把类别用对应编码替换
# print(data)
X = data.iloc[:, 1:-1].values # 第一个‘：’不包括第一列，‘1：-1‘表示从第二列开始，最后一个不要
Y = data.Species.values
train_x, test_x, train_y, test_y = train_test_split(X, Y) # 分训练和测试集
train_x = torch.from_numpy(train_x).type(torch.float32)
test_x = torch.from_numpy(test_x).type(torch.float32)
train_y = torch.from_numpy(train_y).type(torch.int64)
test_y = torch.from_numpy(test_y).type(torch.int64)
batch = 8
train_ds = TensorDataset(train_x, train_y)
train_dl = DataLoader(train_ds, batch_size=batch, shuffle=True)
test_ds = TensorDataset(test_x, test_y)
test_dl = DataLoader(test_ds, batch_size=batch)
class Model(nn.Module):
# 定义一个类，从nn.Module中继承
def __init__(self): # 初始化
super().__init__() # 继承父类中所有属性
self.liner_1 = nn.Linear(4, 32) # 输出20层 输出到隐藏层，创建64个隐藏层
self.liner_2 = nn.Linear(32, 32)
self.liner_3 = nn.Linear(32, 3) # 输出层1
def forward(self, input):
# 在forward中使用上面的层，参数为self，对input进行处理
x = F.relu(self.liner_1(input)) # 第一层调用,直接用F调用relu
x = F.relu(self.liner_2(x))
x = self.liner_3(x) #
return x
model = Model()
loss_fn = nn.CrossEntropyLoss()
optim = torch.optim.Adam(model.parameters(), lr=0.0001)
train_loss = []
train_acc = []
test_loss = []
test_acc = []
epochs = 500
for epoch in range(epochs):
epoch_loss, epoch_acc, epoch_test_loss, epoch_test_acc = fit(epoch, model, train_dl, test_dl)
train_loss.append(epoch_loss)
train_acc.append(epoch_acc)
test_loss.append(epoch_test_loss)
test_acc.append(epoch_test_acc)
1850
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
