python实现LSTM

以下是使用Python实现LSTM的基本代码：

import numpy as np

# 定义LSTM类
class LSTM:
    def __init__(self, input_size, hidden_size, output_size):
        self.input_size = input_size
        self.hidden_size = hidden_size
        self.output_size = output_size

        # 初始化权重和偏置
        self.Wf = np.random.randn(hidden_size + input_size, hidden_size)
        self.Wi = np.random.randn(hidden_size + input_size, hidden_size)
        self.Wc = np.random.randn(hidden_size + input_size, hidden_size)
        self.Wo = np.random.randn(hidden_size + input_size, hidden_size)
        self.Wy = np.random.randn(hidden_size, output_size)

        self.bf = np.zeros((1, hidden_size))
        self.bi = np.zeros((1, hidden_size))
        self.bc = np.zeros((1, hidden_size))
        self.bo = np.zeros((1, hidden_size))
        self.by = np.zeros((1, output_size))

    # 定义sigmoid函数和tanh函数
    def sigmoid(self, x):
        return 1 / (1 + np.exp(-x))

    def tanh(self, x):
        return np.tanh(x)

    # 前向传播
    def forward(self, X):
        T = len(X)
        self.h = np.zeros((T + 1, self.hidden_size))
        self.f = np.zeros((T, self.hidden_size))
        self.i = np.zeros((T, self.hidden_size))
        self.c_bar = np.zeros((T, self.hidden_size))
        self.c = np.zeros((T + 1, self.hidden_size))
        self.o = np.zeros((T, self.hidden_size))
        self.y = np.zeros((T, self.output_size))

        for t in range(T):
            # 将输入和上一时刻的隐藏状态连接起来
            concat = np.hstack((self.h[t], X[t]))

            # 计算遗忘门
            self.f[t] = self.sigmoid(np.dot(concat, self.Wf) + self.bf)

            # 计算输入门
            self.i[t] = self.sigmoid(np.dot(concat, self.Wi) + self.bi)

            # 计算候选状态
            self.c_bar[t] = self.tanh(np.dot(concat, self.Wc) + self.bc)

            # 计算当前状态
            self.c[t + 1] = self.f[t] * self.c[t] + self.i[t] * self.c_bar[t]

            # 计算输出门
            self.o[t] = self.sigmoid(np.dot(concat, self.Wo) + self.bo)

            # 计算输出
            self.h[t + 1] = self.o[t] * self.tanh(self.c[t + 1])
            self.y[t] = np.dot(self.h[t + 1], self.Wy) + self.by

        return self.y

# 测试代码
input_size = 10
hidden_size = 20
output_size = 5
T = 50

X = [np.random.randn(input_size) for t in range(T)]
lstm = LSTM(input_size, hidden_size, output_size)
y = lstm.forward(X)
print(y.shape)

上述代码实现了一个基本的LSTM类，并在测试代码中使用随机数据进行了测试。