鸢尾花识别.zip

import torch import pandas as pd from torch import nn, optim import numpy as np import torch.utils.data.dataloader as DataLoader from torch.autograd import Variable import matplotlib.pyplot as plt from main import IrisDataSet from main.Net import ClasifyNet trainPath = "./datasets/iris.csv" testPath = "./datasets/iris_test.csv" epoches = 200 def weights_init(m): classname = m.__class__.__name__ if classname.find('Linear') != -1: torch.nn.init.normal_(m.weight.data, mean=0, std=1) def compute_accuracy(x, y): z = x == y length = z.shape[0] length = torch.tensor(length).float() print("length = {}".format(length)) summ = z.sum(dim=0).float() print("summ = {}".format(summ)) return summ / length if __name__ == '__main__': # net = ClasifyNet() # net.apply(weights_init) net = torch.load("./test.pth") criterion = nn.CrossEntropyLoss(size_average = True) optimizer = optim.Adam(net.parameters(), lr=0.0001) # 数据准备 train_dataset = IrisDataSet(trainPath) train_data_loader = DataLoader.DataLoader(train_dataset, batch_size=150, shuffle=True, num_workers=4) index = 1 num_x = [] num_y = [] for epoch in range(epoches): for i, item in enumerate(train_data_loader): optimizer.zero_grad() data, label = item data = Variable(data.float()) label = Variable(label.long()) y = net(data) y_temp = y.argmax(dim=1) accuracy = compute_accuracy(label, y_temp) print("y = {}".format(accuracy)) loss = criterion(y, label) loss.backward() optimizer.step() if i == 0: print("losss = {}".format(loss.item())) num_x.append(index) index += 1 num_y.append(loss) plt.plot(num_x, num_y) plt.show() torch.save(net, f='./test.pth')