CNN 和 BiGRU 特征融合实现八分类 | PyTorch 代码示例

在训练集、验证集和测试集的txt文件夹中有如上格式的数据集，每条样本前23位为特征值，最后一位为分类标签，标签共有8个类别，现使用CNN和BiGRU模型并行做为特征提取器，其中输入cnn的输入数据需要将其转化为图片格式，输入BiGRU的数据为原始数据的23位特征值，然后将两提取器的结果进行contact特征融合，再使用全连接层对融合的结果进行八分类，使用pytorch写出正确完整的python代码内容：以下是使用PyTorch编写的代码，实现了CNN和BiGRU模型并行作为特征提取器，并使用全连接层进行八分类。\n\npython\nimport torch\nimport torch.nn as nn\nimport torch.optim as optim\nimport torchvision.transforms as transforms\nfrom torch.utils.data import DataLoader, Dataset\nfrom torch.nn.utils.rnn import pack_padded_sequence, pad_packed_sequence\n\n# 定义CNN模型\nclass CNN(nn.Module):\n def __init__(self):\n super(CNN, self).__init__()\n self.conv1 = nn.Conv2d(1, 16, kernel_size=3, stride=1, padding=1)\n self.relu1 = nn.ReLU()\n self.pool1 = nn.MaxPool2d(kernel_size=2, stride=2)\n self.conv2 = nn.Conv2d(16, 32, kernel_size=3, stride=1, padding=1)\n self.relu2 = nn.ReLU()\n self.pool2 = nn.MaxPool2d(kernel_size=2, stride=2)\n self.fc = nn.Linear(32 * 7 * 7, 128)\n\n def forward(self, x):\n x = self.conv1(x)\n x = self.relu1(x)\n x = self.pool1(x)\n x = self.conv2(x)\n x = self.relu2(x)\n x = self.pool2(x)\n x = x.view(x.size(0), -1)\n x = self.fc(x)\n return x\n\n# 定义BiGRU模型\nclass BiGRU(nn.Module):\n def __init__(self, input_size, hidden_size, num_layers):\n super(BiGRU, self).__init__()\n self.gru = nn.GRU(input_size, hidden_size, num_layers, batch_first=True, bidirectional=True)\n self.fc = nn.Linear(hidden_size * 2, 128)\n\n def forward(self, x):\n x, _ = self.gru(x)\n x = self.fc(x[:, -1, :])\n return x\n\n# 定义数据集类\nclass MyDataset(Dataset):\n def __init__(self, data_path):\n self.data = []\n self.labels = []\n with open(data_path, 'r') as f:\n lines = f.readlines()\n for line in lines:\n line = line.strip().split(',')\n features = list(map(float, line[:-1]))\n label = int(line[-1])\n self.data.append(features)\n self.labels.append(label)\n\n def __len__(self):\n return len(self.data)\n\n def __getitem__(self, index):\n features = torch.tensor(self.data[index], dtype=torch.float32)\n label = torch.tensor(self.labels[index], dtype=torch.long)\n return features, label\n\n# 定义模型\nclass Model(nn.Module):\n def __init__(self, cnn, bigru):\n super(Model, self).__init__()\n self.cnn = cnn\n self.bigru = bigru\n self.fc = nn.Linear(256, 8)\n\n def forward(self, x1, x2):\n x1 = x1.unsqueeze(1) # 转为图片格式，增加通道维度\n x1 = self.cnn(x1)\n x2 = self.bigru(x2)\n x = torch.cat((x1, x2), dim=1) # 特征融合\n x = self.fc(x)\n return x\n\n# 定义训练函数\ndef train(model, device, train_loader, optimizer, criterion):\n model.train()\n for batch_idx, (data1, data2) in enumerate(train_loader):\n data1, data2 = data1.to(device), data2.to(device)\n optimizer.zero_grad()\n output = model(data1, data2)\n loss = criterion(output, data2)\n loss.backward()\n optimizer.step()\n\n# 定义测试函数\ndef test(model, device, test_loader):\n model.eval()\n correct = 0\n with torch.no_grad():\n for data1, data2 in test_loader:\n data1, data2 = data1.to(device), data2.to(device)\n output = model(data1, data2)\n pred = output.argmax(dim=1, keepdim=True)\n correct += pred.eq(data2.view_as(pred)).sum().item()\n accuracy = 100. * correct / len(test_loader.dataset)\n return accuracy\n\n# 设置参数\nbatch_size = 32\nepochs = 10\nlearning_rate = 0.001\n\n# 加载数据\ntrain_dataset = MyDataset('train.txt')\ntrain_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\ntest_dataset = MyDataset('test.txt')\ntest_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False)\n\n# 创建模型和优化器\ncnn = CNN()\nbigru = BiGRU(input_size=23, hidden_size=64, num_layers=2)\nmodel = Model(cnn, bigru)\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\nmodel.to(device)\noptimizer = optim.Adam(model.parameters(), lr=learning_rate)\ncriterion = nn.CrossEntropyLoss()\n\n# 训练和测试\nfor epoch in range(epochs):\n train(model, device, train_loader, optimizer, criterion)\n accuracy = test(model, device, test_loader)\n print('Epoch: {}, Accuracy: {:.2f}%'.format(epoch+1, accuracy))\n\n\n请确保将训练集和测试集的数据分别保存为train.txt和test.txt文件，并将这两个文件放在相同的目录下。你还可以根据需要调整CNN和BiGRU模型的结构，以及训练的迭代次数、学习率等超参数。