将10张图片进行节点分类每张图片有20个节点从节点0到节点19每张图66条边储存在CUsersjhDesktopdatarawedgescsv中每行为一条边第1列为源节点第2列为目标节点；1个节点包含2个特征所有图片节点的第一个特征储存在CUsersjhDesktopdatarawfeatures1csv中第二个特征储存在CUsersjhDesktopdatarawfeatures2csv中其中每

下面是根据你的要求编写的代码：

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch_geometric.data import Dataset
from torch_geometric.data import Data

class My_dataset(Dataset):
    def __init__(self, root):
        self.root = root
        self.edges = torch.tensor([], dtype=torch.long)
        self.features1 = torch.tensor([], dtype=torch.float)
        self.features2 = torch.tensor([], dtype=torch.float)
        self.labels = torch.tensor([], dtype=torch.long)
        self.train_mask = torch.tensor([], dtype=torch.bool)
        
        # Load data from files
        for i in range(10):
            edges = torch.tensor([], dtype=torch.long)
            features1 = torch.tensor([], dtype=torch.float)
            features2 = torch.tensor([], dtype=torch.float)
            labels = torch.tensor([], dtype=torch.long)
            
            # Load edges
            edges_file = f"{self.root}/edges_{i}.csv"
            with open(edges_file, 'r') as f:
                lines = f.readlines()
                for line in lines:
                    src, tgt = line.strip().split(',')
                    edges = torch.cat((edges, torch.tensor([[int(src)], [int(tgt)]])), dim=1)

            # Load features1
            features1_file = f"{self.root}/features1_{i}.csv"
            with open(features1_file, 'r') as f:
                lines = f.readlines()
                for line in lines:
                    feats = [float(x) for x in line.strip().split(',')]
                    features1 = torch.cat((features1, torch.tensor([feats])), dim=0)

            # Load features2
            features2_file = f"{self.root}/features2_{i}.csv"
            with open(features2_file, 'r') as f:
                lines = f.readlines()
                for line in lines:
                    feats = [float(x) for x in line.strip().split(',')]
                    features2 = torch.cat((features2, torch.tensor([feats])), dim=0)

            # Load labels
            labels_file = f"{self.root}/labels_{i}.csv"
            with open(labels_file, 'r') as f:
                lines = f.readlines()
                for line in lines:
                    lbls = [int(x) for x in line.strip().split(',')]
                    labels = torch.cat((labels, torch.tensor([lbls])), dim=0)

            # Create train mask
            train_mask = torch.ones(features1.size(0), dtype=torch.bool)
            train_mask[:16] = False

            self.edges = torch.cat((self.edges, edges), dim=1)
            self.features1 = torch.cat((self.features1, features1), dim=0)
            self.features2 = torch.cat((self.features2, features2), dim=0)
            self.labels = torch.cat((self.labels, labels), dim=0)
            self.train_mask = torch.cat((self.train_mask, train_mask), dim=0)

    def __len__(self):
        return self.features1.size(0)

    def __getitem__(self, idx):
        edge_index = self.edges[:, idx].view(2, -1)
        x = torch.cat((self.features1[idx], self.features2[idx]), dim=1)
        y = self.labels[idx]
        train_mask = self.train_mask[idx]
        data = Data(x=x, edge_index=edge_index, y=y, train_mask=train_mask)
        return data


class GCN(nn.Module):
    def __init__(self, input_dim, hidden_dim, output_dim):
        super(GCN, self).__init__()
        self.conv1 = nn.Conv1d(input_dim, hidden_dim, kernel_size=1)
        self.conv2 = nn.Conv1d(hidden_dim, output_dim, kernel_size=1)

    def forward(self, x, edge_index):
        x = x.permute(0, 2, 1)
        x = F.relu(self.conv1(x))
        x = F.dropout(x, training=self.training)
        x = self.conv2(x)
        x = x.permute(0, 2, 1)
        x = F.log_softmax(x, dim=2)
        return x


# Create dataset
dataset = My_dataset("C:/Users/jh/Desktop/data/raw")

# Create GCN model
model = GCN(input_dim=2, hidden_dim=16, output_dim=2)

# Train and validate the model
for data in dataset:
    optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=5e-4)
    criterion = nn.NLLLoss()

    model.train()
    optimizer.zero_grad()
    out = model(data.x, data.edge_index)
    loss = criterion(out[data.train_mask], data.y[data.train_mask])
    loss.backward()
    optimizer.step()

    model.eval()
    with torch.no_grad():
        pred = model(data.x, data.edge_index)
        val_loss = criterion(pred[~data.train_mask], data.y[~data.train_mask])
        pred = pred.argmax(dim=2)
        acc = pred.eq(data.y.view(-1, 1)).sum().item() / pred.numel()

    print(f"Loss: {loss.item()}, Val Loss: {val_loss.item()}, Accuracy: {acc}")

请注意，以上代码假设每个节点的特征维度为2，GCN模型的第一个卷积层将输入的特征维度从2转换为16，第二个卷积层将输入的特征维度从16转换为2。你可以根据实际情况进行调整。此外，你需要将文件路径中的"C:/Users/jh/Desktop/data/raw"修改为你自己的文件路径。