使用PyTorch写一个图神经网络
时间: 2023-10-02 19:10:17 浏览: 108
graph-neural-networks:在PyTorch中实现图神经网络的库
这里提供一个简单的图神经网络的PyTorch实现,用于节点分类任务。
首先,我们需要导入必要的库:
```python
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch_geometric.datasets import Planetoid
import torch_geometric.nn as pyg_nn
import torch_geometric.transforms as T
```
然后,我们需要加载数据集。这里使用了Planetoid数据集中的Cora数据集。
```python
dataset = Planetoid(root='/data', name='Cora', transform=T.NormalizeFeatures())
data = dataset[0]
```
接下来,我们定义一个简单的图神经网络模型。这个模型包含一个GCN层和一个线性层。
```python
class Net(torch.nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = pyg_nn.GCNConv(dataset.num_node_features, 16)
self.fc = torch.nn.Linear(16, dataset.num_classes)
def forward(self, data):
x, edge_index = data.x, data.edge_index
x = self.conv1(x, edge_index)
x = F.relu(x)
x = F.dropout(x, training=self.training)
x = self.fc(x)
return F.log_softmax(x, dim=1)
```
在forward函数中,我们首先将节点特征x和边信息edge_index传入GCN层中,得到GCN层的输出。然后,我们对输出进行ReLU激活和dropout操作,最后将其输入线性层进行分类。
接下来,我们定义训练函数和测试函数。
```python
def train(model, optimizer, data):
model.train()
optimizer.zero_grad()
out = model(data)
loss = F.nll_loss(out[data.train_mask], data.y[data.train_mask])
loss.backward()
optimizer.step()
return loss.item()
def test(model, data):
model.eval()
out = model(data)
pred = out.argmax(dim=1)
acc = pred[data.test_mask].eq(data.y[data.test_mask]).sum().item() / data.test_mask.sum().item()
return acc
```
在训练函数中,我们首先将模型设为训练模式,然后使用优化器将梯度清零。接着,我们将数据传入模型中,得到输出并计算损失。最后,我们反向传播并更新模型参数。
在测试函数中,我们首先将模型设为测试模式,然后将数据传入模型中得到输出。接着,我们将输出转换为预测结果,并计算准确率。
最后,我们开始训练和测试模型。
```python
model = Net()
optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=5e-4)
for epoch in range(200):
loss = train(model, optimizer, data)
acc = test(model, data)
print('Epoch {}, Loss: {:.4f}, Test Acc: {:.4f}'.format(epoch, loss, acc))
```
这里我们使用Adam优化器和交叉熵损失函数,进行200个epoch的训练。每个epoch结束后,我们输出当前的损失和测试准确率。
完整代码如下:
```python
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch_geometric.datasets import Planetoid
import torch_geometric.nn as pyg_nn
import torch_geometric.transforms as T
dataset = Planetoid(root='/data', name='Cora', transform=T.NormalizeFeatures())
data = dataset[0]
class Net(torch.nn.Module):
def __init__(self):
super(Net, self).__init__()
self.conv1 = pyg_nn.GCNConv(dataset.num_node_features, 16)
self.fc = torch.nn.Linear(16, dataset.num_classes)
def forward(self, data):
x, edge_index = data.x, data.edge_index
x = self.conv1(x, edge_index)
x = F.relu(x)
x = F.dropout(x, training=self.training)
x = self.fc(x)
return F.log_softmax(x, dim=1)
def train(model, optimizer, data):
model.train()
optimizer.zero_grad()
out = model(data)
loss = F.nll_loss(out[data.train_mask], data.y[data.train_mask])
loss.backward()
optimizer.step()
return loss.item()
def test(model, data):
model.eval()
out = model(data)
pred = out.argmax(dim=1)
acc = pred[data.test_mask].eq(data.y[data.test_mask]).sum().item() / data.test_mask.sum().item()
return acc
model = Net()
optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=5e-4)
for epoch in range(200):
loss = train(model, optimizer, data)
acc = test(model, data)
print('Epoch {}, Loss: {:.4f}, Test Acc: {:.4f}'.format(epoch, loss, acc))
```
阅读全文