英译法案例 - 测试基于 GRU 的无 Attention 解码器

import torch
import torch.nn as nn
import torch.nn.functional as F


class EncoderGRU(nn.Module):
    """基于 GRU 的编码器"""
    
    def __init__(self, vocab_size, hidden_size):
        super(EncoderGRU, self).__init__()
        self.hidden_size = hidden_size
        self.embedding = nn.Embedding(vocab_size, hidden_size)
        self.gru = nn.GRU(hidden_size, hidden_size)
    
    def forward(self, input, hidden):
        embedded = self.embedding(input).view(1, 1, -1)
        output, hidden = self.gru(embedded, hidden)
        return output, hidden
    
    def initHidden(self):
        return torch.zeros(1, 1, self.hidden_size)


class DecoderGRU(nn.Module):
    """基于 GRU 的无 Attention 解码器"""
    
    def __init__(self, hidden_size, output_size):
        super(DecoderGRU, self).__init__()
        self.hidden_size = hidden_size
        
        # 1. 词嵌入层：4345 → 256
        self.embedding = nn.Embedding(output_size, hidden_size)
        
        # 2. GRU 层：256 → 256
        self.gru = nn.GRU(hidden_size, hidden_size)
        
        # 3. 线性层：256 → 4345
        self.out = nn.Linear(hidden_size, output_size)
        
        # 4. LogSoftmax
        self.softmax = nn.LogSoftmax(dim=1)
    
    def forward(self, input, hidden):
        embedded = self.embedding(input).view(1, 1, -1)
        embedded = F.relu(embedded)
        output, hidden = self.gru(embedded, hidden)
        output = self.softmax(self.out(output[0]))
        return output, hidden
    
    def initHidden(self):
        return torch.zeros(1, 1, self.hidden_size)


def test_decoder_gru():
    """测试基于 GRU 的无 Attention 解码器"""
    print("=== 英译法案例 - 测试基于 GRU 的无 Attention 解码器 ===")
    
    # 1. 获取数据加载器对象
    # my_data_loader = get_data_loader()
    
    # 2. 初始化编码器模型并移动到 GPU
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    my_encoder_gru = EncoderGRU(vocab_size=28036, hidden_size=256).to(device)
    
    # 3. 初始化解码器模型并移动到 GPU
    my_decoder_gru = DecoderGRU(hidden_size=256, output_size=4345).to(device)
    
    print(f"设备: {device}")
    print(f"\n编码器:\n{my_encoder_gru}")
    print(f"\n解码器:\n{my_decoder_gru}")
    
    # 4. 测试
    input_tensor = torch.tensor([[1]]).to(device)  # [1, 1]
    
    # 编码器
    encoder_hidden = my_encoder_gru.initHidden().to(device)
    encoder_output, encoder_hidden = my_encoder_gru(input_tensor, encoder_hidden)
    
    # 解码器
    decoder_input = torch.tensor([[1]]).to(device)  # SOS
    decoder_hidden = encoder_hidden  # 编码器的隐藏状态作为解码器的初始隐藏状态
    
    decoder_output, decoder_hidden = my_decoder_gru(decoder_input, decoder_hidden)
    
    print(f"\n编码器输出形状: {encoder_output.shape}")
    print(f"解码器输出形状: {decoder_output.shape}")  # [1, 4345]
    print(f"概率最高的词索引: {decoder_output.topk(1).indices.item()}")


# 测试
if __name__ == "__main__":
    test_decoder_gru()