task03--循环神经网络进阶

循环神经网络进阶

1、GRU

RNN存在的问题：梯度较容易出现衰减或爆炸（BPTT）
⻔控循环神经⽹络：捕捉时间序列中时间步距离较⼤的依赖关系
RNN:

$$H_t=\varphi (X_t{W}_{xh}+H_{t-1}{W}_{hh}+b_h)$$
GRU:

$$\begin{gathered} R_t=\sigma (X_t{W}_{xr}+H_{t-1}{W}_{hr}+b_r) \\ {Z}_t=\sigma (X_t{W}_{xz}+H_{t-1}{W}_{hz}+b_z) \\ {\tilde{H}}_t=tanh(X_t{W}_{xh}+(R_t\odot H_{t-1})W_{hh}+b_h) \\ {H}_t=Z_t\odot H_{t-1}+(1-Z_t)\odot {\tilde{H}}_t \end{gathered}$$
（Rt 、Zt 、 的形状都是h，Wxr Wxz 的形状是Xh，Whr Whz 的形状是hh）

• 重置⻔有助于捕捉时间序列⾥短期的依赖关系；
• 更新⻔有助于捕捉时间序列⾥⻓期的依赖关系。

def gru(inputs, state, params):
    W_xz, W_hz, b_z, W_xr, W_hr, b_r, W_xh, W_hh, b_h, W_hq, b_q = params
    H, = state
    outputs = []
    for X in inputs:
        Z = torch.sigmoid(torch.matmul(X, W_xz) + torch.matmul(H, W_hz) + b_z)
        R = torch.sigmoid(torch.matmul(X, W_xr) + torch.matmul(H, W_hr) + b_r)
        H_tilda = torch.tanh(torch.matmul(X, W_xh) + R * torch.matmul(H, W_hh) + b_h)
        H = Z * H + (1 - Z) * H_tilda
        Y = torch.matmul(H, W_hq) + b_q
        outputs.append(Y)
    return outputs, (H,)

2、LSTM

长短期记忆long short-term memory
遗忘门:控制上一时间步的记忆细胞
输入门:控制当前时间步的输入
输出门:控制从记忆细胞到隐藏状态
记忆细胞：⼀种特殊的隐藏状态的信息的流动

$$\begin{gathered} I_t=\sigma (X_t{W}_{xi}+H_{t-1}{W}_{hi}+b_i) \\ {F}_t=\sigma (X_t{W}_{xf}+H_{t-1}{W}_{hf}+b_f) \\ {O}_t=\sigma (X_t{W}_{xo}+H_{t-1}{W}_{ho}+b_o) \\ {\tilde{C}}_t=tanh(X_t{W}_{xc}+H_{t-1}{W}_{hc}+b_c) \\ {C}_t=F_t\odot C_{t-1}+I_t\odot {\tilde{C}}_t \\ {H}_t=O_t\odot tanh(C_t) \end{gathered}$$

def lstm(inputs, state, params):
    [W_xi, W_hi, b_i, W_xf, W_hf, b_f, W_xo, W_ho, b_o, W_xc, W_hc, b_c, W_hq, b_q] = params
    (H, C) = state
    outputs = []
    for X in inputs:
        I = torch.sigmoid(torch.matmul(X, W_xi) + torch.matmul(H, W_hi) + b_i)
        F = torch.sigmoid(torch.matmul(X, W_xf) + torch.matmul(H, W_hf) + b_f)
        O = torch.sigmoid(torch.matmul(X, W_xo) + torch.matmul(H, W_ho) + b_o)
        C_tilda = torch.tanh(torch.matmul(X, W_xc) + torch.matmul(H, W_hc) + b_c)
        C = F * C + I * C_tilda
        H = O * C.tanh()
        Y = torch.matmul(H, W_hq) + b_q
        outputs.append(Y)
    return outputs, (H, C)

3、深度循环神经网络

$$\begin{gathered} {\mathbf{H}}_t^{(1)}=\varphi ({\mathbf{X}}_t{\mathbf{W}}_{xh}^{(1)}+{\mathbf{H}}_{t-1}^{(1)}{\mathbf{W}}_{hh}^{(1)}+{\mathbf{b}}_h^{(1)}) \\ {\mathbf{H}}_t^{(\ell )}=\varphi ({\mathbf{H}}_t^{(\ell -1)}{\mathbf{W}}_{xh}^{(\ell )}+{\mathbf{H}}_{t-1}^{(\ell )}{\mathbf{W}}_{hh}^{(\ell )}+{\mathbf{b}}_h^{(\ell )}) \\ {\mathbf{O}}_t={\mathbf{H}}_t^{(L)}{\mathbf{W}}_{hq}+{\mathbf{b}}_q \end{gathered}$$

• 可抽取更高、更深层的语义信息
• 但是层数越深，效果不一定越好，因为模型的复杂度会提升

num_hiddens=256
num_epochs, num_steps, batch_size, lr, clipping_theta = 160, 35, 32, 1e2, 1e-2
pred_period, pred_len, prefixes = 40, 50, ['分开', '不分开']

lr = 1e-2 # 注意调整学习率

gru_layer = nn.LSTM(input_size=vocab_size, hidden_size=num_hiddens,num_layers=2)    #参数num_layers=2，深度为2.默认为false，则是单层的lstm
model = d2l.RNNModel(gru_layer, vocab_size).to(device)
d2l.train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
                                corpus_indices, idx_to_char, char_to_idx,
                                num_epochs, num_steps, lr, clipping_theta,
                                batch_size, pred_period, pred_len, prefixes)

def train_and_predict_rnn_pytorch(model, num_hiddens, vocab_size, device,
                                corpus_indices, idx_to_char, char_to_idx,
                                num_epochs, num_steps, lr, clipping_theta,
                                batch_size, pred_period, pred_len, prefixes):
    loss = nn.CrossEntropyLoss()
    optimizer = torch.optim.Adam(model.parameters(), lr=lr)
    model.to(device)
    for epoch in range(num_epochs):
        l_sum, n, start = 0.0, 0, time.time()
        data_iter = d2l.data_iter_consecutive(corpus_indices, batch_size, num_steps, device) # 相邻采样
        state = None
        for X, Y in data_iter:
            if state is not None:
                # 使用detach函数从计算图分离隐藏状态
                if isinstance (state, tuple): # LSTM, state:(h, c)  
                    state[0].detach_()
                    state[1].detach_()
                else: 
                    state.detach_()
            (output, state) = model(X, state) # output.shape: (num_steps * batch_size, vocab_size)
            y = torch.flatten(Y.T)
            l = loss(output, y.long())
            
            optimizer.zero_grad()
            l.backward()
            grad_clipping(model.parameters(), clipping_theta, device)
            optimizer.step()
            l_sum += l.item() * y.shape[0]
            n += y.shape[0]
        

        if (epoch + 1) % pred_period == 0:
            print('epoch %d, perplexity %f, time %.2f sec' % (
                epoch + 1, math.exp(l_sum / n), time.time() - start))
            for prefix in prefixes:
                print(' -', predict_rnn_pytorch(
                    prefix, pred_len, model, vocab_size, device, idx_to_char,
                    char_to_idx))

class RNNModel(nn.Module):
    def __init__(self, rnn_layer, vocab_size):
        super(RNNModel, self).__init__()
        self.rnn = rnn_layer
        self.hidden_size = rnn_layer.hidden_size * (2 if rnn_layer.bidirectional else 1) 
        self.vocab_size = vocab_size
        self.dense = nn.Linear(self.hidden_size, vocab_size)

    def forward(self, inputs, state):
        # inputs.shape: (batch_size, num_steps)
        X = to_onehot(inputs, vocab_size)
        X = torch.stack(X)  # X.shape: (num_steps, batch_size, vocab_size)
        hiddens, state = self.rnn(X, state)
        hiddens = hiddens.view(-1, hiddens.shape[-1])  # hiddens.shape: (num_steps * batch_size, hidden_size)
        output = self.dense(hiddens)
        return output, state

4.双向循环神经网络

• 可以获取词上下文的信息
• 前向的H_t 和后向的H_t 用concat进行连结

gru_layer = nn.GRU(input_size=vocab_size, hidden_size=num_hiddens,bidirectional=True)