用python torch：使用iris.csv数据构建神经网络预测花的种类：将所有数据随机分为训练集和测试集，分别计算测试集上总的准确率和不同类的准确率。

好的，首先需要导入必要的库和数据集：

import torch
import torch.nn as nn
import pandas as pd
from sklearn.model_selection import train_test_split

# 读取数据集
iris = pd.read_csv('iris.csv')

# 将类别转换为数字
iris['Species'] = pd.Categorical(iris['Species'])
iris['Species'] = iris['Species'].cat.codes

# 划分训练集和测试集
X_train, X_test, y_train, y_test = train_test_split(iris.iloc[:,:-1].values, iris.iloc[:,-1].values, test_size=0.2, random_state=42)

接下来，我们需要定义一个三层的神经网络模型：

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.fc1 = nn.Linear(4, 16)
        self.fc2 = nn.Linear(16, 8)
        self.fc3 = nn.Linear(8, 3)

    def forward(self, x):
        x = torch.relu(self.fc1(x))
        x = torch.relu(self.fc2(x))
        x = self.fc3(x)
        return x

接着，我们需要定义一些训练参数：

net = Net()
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(net.parameters(), lr=0.01)

然后，我们可以开始训练模型：

# 设置训练轮数
epochs = 1000

for epoch in range(epochs):
    optimizer.zero_grad()
    inputs = torch.tensor(X_train, dtype=torch.float32)
    labels = torch.tensor(y_train, dtype=torch.long)
    outputs = net(inputs)
    loss = criterion(outputs, labels)
    loss.backward()
    optimizer.step()

    # 计算训练集上的准确率
    with torch.no_grad():
        inputs = torch.tensor(X_train, dtype=torch.float32)
        labels = torch.tensor(y_train, dtype=torch.long)
        outputs = net(inputs)
        _, predicted = torch.max(outputs.data, 1)
        correct = (predicted == labels).sum().item()
        train_acc = correct / len(y_train)

    # 计算测试集上的准确率
    with torch.no_grad():
        inputs = torch.tensor(X_test, dtype=torch.float32)
        labels = torch.tensor(y_test, dtype=torch.long)
        outputs = net(inputs)
        _, predicted = torch.max(outputs.data, 1)
        correct = (predicted == labels).sum().item()
        test_acc = correct / len(y_test)

    # 输出训练进度
    if (epoch+1) % 100 == 0:
        print('Epoch [{}/{}], Loss: {:.4f}, Train Acc: {:.2f}%, Test Acc: {:.2f}%'
              .format(epoch+1, epochs, loss.item(), train_acc*100, test_acc*100))

最后，我们可以输出不同类别的准确率：

with torch.no_grad():
    inputs = torch.tensor(X_test, dtype=torch.float32)
    labels = torch.tensor(y_test, dtype=torch.long)
    outputs = net(inputs)
    _, predicted = torch.max(outputs.data, 1)
    for i in range(3):
        indices = (labels == i)
        class_acc = (predicted[indices] == labels[indices]).float().mean()
        print('Accuracy of class %d: %.2f%%' % (i, class_acc*100))

完整代码如下：