PytorchLSTM时间序列预测

Pytorch  LSTM 时间序列预测

https://github.com/pytorch/examples/blob/master/time_sequence_prediction/generate_sine_wave.py

Pytorch官网提供初学者入门的一个例子&＃xff0c;有助于学习Pytorch时间序列预测。本例中使用两个LSTMCell单元学习从不同相位开始的一些正弦波信号,LSTM网络在学习了正弦波之后&＃xff0c;试图预测未来的信号值。

generate_sine_wave.py生成模拟数据&＃xff1a;

# -*- coding: utf-8 -*-
import numpy as np
import torch
np.random.seed(2)
T &＃61; 20
L &＃61; 1000
N &＃61; 100
x &＃61; np.empty((N, L), &＃39;int64&＃39;)
x[:] &＃61; np.array(range(L)) &＃43; np.random.randint(-4 * T, 4 * T, N).reshape(N, 1)
data &＃61; np.sin(x / 1.0 / T).astype(&＃39;float64&＃39;)
torch.save(data, open(&＃39;traindata.pt&＃39;, &＃39;wb&＃39;))


LSTM数据序列预测&＃xff1a;

# -*- coding: utf-8 -*-
from __future__ import print_function
import torch
import torch.nn as nn
import torch.optim as optim
import numpy a