import torch import torch.nn as nn import torch.optim as optim from torch.utils.data import Dataset, DataLoader import pandas as pd # 1. 数据集类 class NameDataset(Dataset): def __init__(self, file_path): data = pd.read_excel(file_path, engine='openpyxl') self.names = data['姓名'].tolist() self.genders = data['性别'].map({'男': 0, '女': 1}).tolist() # 构建字符字典 all_chars = ''.join(self.names) self.chars = sorted(list(set(all_chars))) self.char2idx = {char: idx for idx, char in enumerate(self.chars)} self.vocab_size = len(self.chars) # 计算最大姓名长度 self.max_length = max(len(name) for name in self.names) def __len__(self): return len(self.names) def __getitem__(self, idx): name = self.names[idx] gender = self.genders[idx] # 字符转索引并填充到固定长度 char_indices = [self.char2idx[char] for char in name] padded = torch.nn.functional.pad( torch.tensor(char_indices, dtype=torch.long), (0, self.max_length - len(char_indices)), value=self.char2idx[' '] # 使用空格作为填充符 ) return padded, torch.tensor(gender, dtype=torch.long) # 2. 模型定义 class NameGenderCNN(nn.Module): def __init__(self, vocab_size, embed_dim=32, hidden_dim=64): super().__init__() self.embedding = nn.Embedding(vocab_size, embed_dim) self.conv1 = nn.Conv1d(embed_dim, hidden_dim, kernel_size=3, padding=1) self.conv2 = nn.Conv1d(hidden_dim, hidden_dim, kernel_size=3, padding=1) self.fc = nn.Linear(hidden_dim, 2) def forward(self, x): x = self.embedding(x).permute(0, 2, 1) # (batch, embed, length) x = nn.functional.relu(self.conv1(x)) x = nn.functional.relu(self.conv2(x)) x = x.mean(dim=2) # 全局平均池化 return self.fc(x) # 3. 训练流程 def train_model(): # 初始化数据集 dataset = NameDataset("D:\\名单数据库.xlsx") dataloader = DataLoader(dataset, batch_size=8, shuffle=True) # 初始化模型 model = NameGenderCNN(vocab_size=dataset.vocab_size) criterion = nn.CrossEntropyLoss() optimizer = optim.Adam(model.parameters(), lr=0.001) # 训练循环 for epoch in range(50): model.train() total_loss = 0 correct = 0 total = 0 for batch_x, batch_y in dataloader: optimizer.zero_grad() outputs = model(batch_x) loss = criterion(outputs, batch_y) loss.backward() optimizer.step() total_loss += loss.item() _, predicted = torch.max(outputs.data, 1) total += batch_y.size(0) correct += (predicted == batch_y).sum().item() accuracy = correct / total print(f'Epoch {epoch + 1}, Loss: {total_loss / len(dataloader):.4f}, Accuracy: {accuracy:.4f}') # 保存模型 torch.save(model.state_dict(), 'gender_model.pth') print("Model saved successfully!") # 4. 预测函数 def predict_gender(name, model_path='gender_model.pth'): model = NameGenderCNN(vocab_size=dataset.vocab_size) model.load_state_dict(torch.load(model_path)) model.eval() dataset = NameDataset("D:\\名单数据库.xlsx") char_indices = [dataset.char2idx.get(c, dataset.char2idx[' ']) for c in name] padded = torch.nn.functional.pad( torch.tensor(char_indices, dtype=torch.long), (0, dataset.max_length - len(char_indices)), value=dataset.char2idx[' '] ).unsqueeze(0) with torch.no_grad(): output = model(padded) _, predicted = torch.max(output, 1) return '男' if predicted.item() == 0 else '女' # 执行训练 train_model() # 示例预测 print(predict_gender("陈光健")) # 男 print(predict_gender("梁慧婷")) # 女 print(predict_gender("张鑫")) # 男 print(predict_gender("王珈桐")) # 女 将上述代码修改正确后给我

<< 以下是对原始代码中的潜在问题进行修正后的完整版本： ### 主要改动点说明： 1. **`NameDataset` 类中 `char2idx[' ']` 错误**： 如果输入数据集中没有空格，则 `char2idx` 中可能不存在键 `' '`，这会导致报错。因此，在构建字符字典时显式添加了空格。 2. **预测部分加载 `dataset` 冗余**： 在预测阶段无需重新初始化整个数据集，只需复用训练时生成的字符映射表即可。 3. **模型路径和文件读取路径改进**： 将硬编码路径替换为可配置参数，方便切换不同环境下的运行需求。 --- #### 修改后的完整代码 ```python import torch import torch.nn as nn import torch.optim as optim from torch.utils.data import Dataset, DataLoader import pandas as pd # 1. 数据集类 class NameDataset(Dataset): def __init__(self, file_path): data = pd.read_excel(file_path, engine='openpyxl') # 确保数据列存在 if '姓名' not in data.columns or '性别' not in data.columns: raise ValueError("Excel 文件必须包含 '姓名' 和 '性别' 列") self.names = data['姓名'].tolist() self.genders = data['性别'].map({'男': 0, '女': 1}).fillna(-1).astype(int).tolist() # 构建字符字典（包括空格） all_chars = ''.join(self.names) + ' ' unique_chars = sorted(list(set(all_chars))) self.chars = ['<PAD>'] + unique_chars # 添加特殊标记 <PAD> self.char2idx = {char: idx for idx, char in enumerate(self.chars)} self.vocab_size = len(self.chars) # 计算最大姓名长度 self.max_length = max(len(name) for name in self.names) def __len__(self): return len(self.names) def __getitem__(self, idx): name = self.names[idx] gender = self.genders[idx] # 字符转索引并填充到固定长度 char_indices = [self.char2idx[char] for char in name if char in self.char2idx] padded = torch.nn.functional.pad( torch.tensor(char_indices, dtype=torch.long), (0, self.max_length - len(char_indices)), value=self.char2idx['<PAD>'] ) return padded, torch.tensor(gender, dtype=torch.long) # 2. 模型定义 class NameGenderCNN(nn.Module): def __init__(self, vocab_size, embed_dim=32, hidden_dim=64): super(NameGenderCNN, self).__init__() self.embedding = nn.Embedding(vocab_size, embed_dim) self.conv1 = nn.Conv1d(embed_dim, hidden_dim, kernel_size=3, padding=1) self.conv2 = nn.Conv1d(hidden_dim, hidden_dim, kernel_size=3, padding=1) self.fc = nn.Linear(hidden_dim, 2) def forward(self, x): x = self.embedding(x).permute(0, 2, 1) # (batch, embed, length) x = nn.functional.relu(self.conv1(x)) x = nn.functional.relu(self.conv2(x)) x = x.mean(dim=2) # 全局平均池化 return self.fc(x) # 3. 训练流程 def train_model(data_file="D:\\名单数据库.xlsx", model_save_path="gender_model.pth"): # 初始化数据集 dataset = NameDataset(data_file) dataloader = DataLoader(dataset, batch_size=8, shuffle=True) # 初始化模型 model = NameGenderCNN(vocab_size=dataset.vocab_size) criterion = nn.CrossEntropyLoss() optimizer = optim.Adam(model.parameters(), lr=0.001) # 训练循环 for epoch in range(50): model.train() total_loss = 0 correct = 0 total = 0 for batch_x, batch_y in dataloader: optimizer.zero_grad() outputs = model(batch_x) loss = criterion(outputs, batch_y) loss.backward() optimizer.step() total_loss += loss.item() _, predicted = torch.max(outputs.data, 1) total += batch_y.size(0) correct += (predicted == batch_y).sum().item() accuracy = correct / total print(f"Epoch {epoch + 1}, Loss: {total_loss / len(dataloader):.4f}, Accuracy: {accuracy:.4f}") # 保存模型 torch.save(model.state_dict(), model_save_path) print("Model saved successfully!") # 4. 预测函数 def predict_gender(name, model_path='gender_model.pth', dataset=None): if dataset is None: raise ValueError("predict_gender 函数需要传入已知的 dataset 对象以获取词汇表信息") model = NameGenderCNN(vocab_size=dataset.vocab_size) model.load_state_dict(torch.load(model_path)) model.eval() # 名称转换为索引序列 char_indices = [ dataset.char2idx.get(c, dataset.char2idx['<PAD>']) for c in name[:dataset.max_length] ] padded = torch.nn.functional.pad( torch.tensor(char_indices, dtype=torch.long), (0, dataset.max_length - len(char_indices)), value=dataset.char2idx['<PAD>'] ).unsqueeze(0) with torch.no_grad(): output = model(padded) _, predicted = torch.max(output, 1) return '男' if predicted.item() == 0 else '女' # 执行训练 if __name__ == "__main__": train_data_file = "D:\\名单数据库.xlsx" trained_model_path = "gender_model.pth" # 步骤一：训练模型 train_dataset = NameDataset(train_data_file) train_model(data_file=train_data_file, model_save_path=trained_model_path) # 步骤二：测试预测功能 test_names = ["陈光健", "梁慧婷", "张鑫", "王珈桐"] predictions = [(name, predict_gender(name, model_path=trained_model_path, dataset=train_dataset)) for name in test_names] for name, pred in predictions: print(f"{name} -> {pred}") ``` --- #### 测试结果示例输出： 假设经过充分训练后： ``` 陈光健 -> 男 梁慧婷 -> 女 张鑫 -> 男 王珈桐 -> 女 ``` ---