• 🍨 本文為🔗365天深度學習訓練營 中的學習記錄博客
• 🍖 原作者：K同學啊

一、前期準備工作

import torch.nn.functional as F
import numpy  as np
import pandas as pd
import torch
from torch    import nn

1.導入數據

data = pd.read_csv("woodpine2.csv")data

	Time	Tem1	CO 1	Soot 1
0	0.000	25.0	0.000000	0.000000
1	0.228	25.0	0.000000	0.000000
2	0.456	25.0	0.000000	0.000000
3	0.685	25.0	0.000000	0.000000
4	0.913	25.0	0.000000	0.000000
...	...	...	...	...
5943	366.000	295.0	0.000077	0.000496
5944	366.000	294.0	0.000077	0.000494
5945	367.000	292.0	0.000077	0.000491
5946	367.000	291.0	0.000076	0.000489
5947	367.000	290.0	0.000076	0.000487

5948 rows × 4 columns

2. 數據集可視化

import matplotlib.pyplot as plt
import seaborn as snsplt.rcParams['savefig.dpi'] = 500 #圖片像素
plt.rcParams['figure.dpi']  = 500 #分辨率fig, ax =plt.subplots(1,3,constrained_layout=True, figsize=(14, 3))sns.lineplot(data=data["Tem1"], ax=ax[0])
sns.lineplot(data=data["CO 1"], ax=ax[1])
sns.lineplot(data=data["Soot 1"], ax=ax[2])
plt.show()

dataFrame = data.iloc[:,1:]dataFrame

	Tem1	CO 1	Soot 1
0	25.0	0.000000	0.000000
1	25.0	0.000000	0.000000
2	25.0	0.000000	0.000000
3	25.0	0.000000	0.000000
4	25.0	0.000000	0.000000
...	...	...	...
5943	295.0	0.000077	0.000496
5944	294.0	0.000077	0.000494
5945	292.0	0.000077	0.000491
5946	291.0	0.000076	0.000489
5947	290.0	0.000076	0.000487

5948 rows × 3 columns

二、構建數據集

1. 數據集預處理

from sklearn.preprocessing import MinMaxScalerdataFrame = data.iloc[:,1:].copy()
sc  = MinMaxScaler(feature_range=(0, 1)) #將數據歸一化，范圍是0到1for i in ['CO 1', 'Soot 1', 'Tem1']:dataFrame[i] = sc.fit_transform(dataFrame[i].values.reshape(-1, 1))dataFrame.shape

(5948, 3)

2. 設置X, y

width_X = 8
width_y = 1##取前8個時間段的Tem1、CO 1、Soot 1為X，第9個時間段的Tem1為y。
X = []
y = []in_start = 0for _, _ in data.iterrows():in_end  = in_start + width_Xout_end = in_end   + width_yif out_end < len(dataFrame):X_ = np.array(dataFrame.iloc[in_start:in_end , ])y_ = np.array(dataFrame.iloc[in_end  :out_end, 0])X.append(X_)y.append(y_)in_start += 1X = np.array(X)
y = np.array(y).reshape(-1,1,1)X.shape, y.shape

((5939, 8, 3), (5939, 1, 1))

檢查數據集中是否有空值

print(np.any(np.isnan(X)))
print(np.any(np.isnan(y)))

False
False

3. 劃分數據集

X_train = torch.tensor(np.array(X[:5000]), dtype=torch.float32)
y_train = torch.tensor(np.array(y[:5000]), dtype=torch.float32)X_test  = torch.tensor(np.array(X[5000:]), dtype=torch.float32)
y_test  = torch.tensor(np.array(y[5000:]), dtype=torch.float32)
X_train.shape, y_train.shape

(torch.Size([5000, 8, 3]), torch.Size([5000, 1, 1]))

from torch.utils.data import TensorDataset, DataLoadertrain_dl = DataLoader(TensorDataset(X_train, y_train),batch_size=64, shuffle=False)test_dl  = DataLoader(TensorDataset(X_test, y_test),batch_size=64, shuffle=False)

三、模型訓練

1. 構建模型

class model_lstm(nn.Module):def __init__(self):super(model_lstm, self).__init__()self.lstm0 = nn.LSTM(input_size=3 ,hidden_size=320, num_layers=1, batch_first=True)self.lstm1 = nn.LSTM(input_size=320 ,hidden_size=320, num_layers=1, batch_first=True)self.fc0   = nn.Linear(320, 1)def forward(self, x):out, hidden1 = self.lstm0(x) out, _ = self.lstm1(out, hidden1) out    = self.fc0(out) return out[:, -1:, :]   #取1個預測值,否則經過lstm會得到8*1個預測model = model_lstm()
model

model_lstm((lstm0): LSTM(3, 320, batch_first=True)(lstm1): LSTM(320, 320, batch_first=True)(fc0): Linear(in_features=320, out_features=1, bias=True)
)

model(torch.rand(30,8,3)).shape

torch.Size([30, 1, 1])

2. 定義訓練函數

# 訓練循環
import copy
def train(train_dl, model, loss_fn, opt, lr_scheduler=None):size        = len(train_dl.dataset)  num_batches = len(train_dl)   train_loss  = 0  # 初始化訓練損失和正確率for x, y in train_dl:  x, y = x.to(device), y.to(device)# 計算預測誤差pred = model(x)          # 網絡輸出loss = loss_fn(pred, y)  # 計算網絡輸出和真實值之間的差距# 反向傳播opt.zero_grad()  # grad屬性歸零loss.backward()  # 反向傳播opt.step()       # 每一步自動更新# 記錄losstrain_loss += loss.item()if lr_scheduler is not None:lr_scheduler.step()print("learning rate = {:.5f}".format(opt.param_groups[0]['lr']), end="  ")train_loss /= num_batchesreturn train_loss

3. 定義測試函數

def test (dataloader, model, loss_fn):size        = len(dataloader.dataset)  # 測試集的大小num_batches = len(dataloader)          # 批次數目test_loss   = 0# 當不進行訓練時，停止梯度更新，節省計算內存消耗with torch.no_grad():for x, y in dataloader:x, y = x.to(device), y.to(device)# 計算lossy_pred = model(x)loss        = loss_fn(y_pred, y)test_loss += loss.item()test_loss /= num_batchesreturn test_loss

4. 正式訓練模型

#設置GPU訓練
device=torch.device("cuda" if torch.cuda.is_available() else "cpu")
device

device(type=‘cpu’)

#訓練模型
model = model_lstm()
model = model.to(device)
loss_fn    = nn.MSELoss() # 創建損失函數
learn_rate = 1e-1   # 學習率
opt        = torch.optim.SGD(model.parameters(),lr=learn_rate,weight_decay=1e-4)
epochs     = 50
train_loss = []
test_loss  = []
lr_scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(opt,epochs, last_epoch=-1) for epoch in range(epochs):model.train()epoch_train_loss = train(train_dl, model, loss_fn, opt, lr_scheduler)model.eval()epoch_test_loss = test(test_dl, model, loss_fn)train_loss.append(epoch_train_loss)test_loss.append(epoch_test_loss)template = ('Epoch:{:2d}, Train_loss:{:.5f}, Test_loss:{:.5f}')print(template.format(epoch+1, epoch_train_loss,  epoch_test_loss))print("="*20, 'Done', "="*20)

learning rate = 0.09990 Epoch: 1, Train_loss:0.00123, Test_loss:0.01228
learning rate = 0.09961 Epoch: 2, Train_loss:0.01404, Test_loss:0.01183
learning rate = 0.09911 Epoch: 3, Train_loss:0.01365, Test_loss:0.01135
learning rate = 0.09843 Epoch: 4, Train_loss:0.01321, Test_loss:0.01085
learning rate = 0.09755 Epoch: 5, Train_loss:0.01270, Test_loss:0.01029
learning rate = 0.09649 Epoch: 6, Train_loss:0.01212, Test_loss:0.00968
learning rate = 0.09524 Epoch: 7, Train_loss:0.01144, Test_loss:0.00901
learning rate = 0.09382 Epoch: 8, Train_loss:0.01065, Test_loss:0.00827
learning rate = 0.09222 Epoch: 9, Train_loss:0.00975, Test_loss:0.00748
learning rate = 0.09045 Epoch:10, Train_loss:0.00876, Test_loss:0.00665
learning rate = 0.08853 Epoch:11, Train_loss:0.00769, Test_loss:0.00580
learning rate = 0.08645 Epoch:12, Train_loss:0.00658, Test_loss:0.00497
learning rate = 0.08423 Epoch:13, Train_loss:0.00548, Test_loss:0.00418
learning rate = 0.08187 Epoch:14, Train_loss:0.00444, Test_loss:0.00346
learning rate = 0.07939 Epoch:15, Train_loss:0.00349, Test_loss:0.00283
learning rate = 0.07679 Epoch:16, Train_loss:0.00268, Test_loss:0.00230
learning rate = 0.07409 Epoch:17, Train_loss:0.00200, Test_loss:0.00188
learning rate = 0.07129 Epoch:18, Train_loss:0.00147, Test_loss:0.00154
learning rate = 0.06841 Epoch:19, Train_loss:0.00107, Test_loss:0.00129
learning rate = 0.06545 Epoch:20, Train_loss:0.00078, Test_loss:0.00110
learning rate = 0.06243 Epoch:21, Train_loss:0.00057, Test_loss:0.00096
learning rate = 0.05937 Epoch:22, Train_loss:0.00042, Test_loss:0.00085
learning rate = 0.05627 Epoch:23, Train_loss:0.00032, Test_loss:0.00078
learning rate = 0.05314 Epoch:24, Train_loss:0.00025, Test_loss:0.00072
learning rate = 0.05000 Epoch:25, Train_loss:0.00021, Test_loss:0.00068
learning rate = 0.04686 Epoch:26, Train_loss:0.00017, Test_loss:0.00065
learning rate = 0.04373 Epoch:27, Train_loss:0.00015, Test_loss:0.00062
learning rate = 0.04063 Epoch:28, Train_loss:0.00014, Test_loss:0.00060
learning rate = 0.03757 Epoch:29, Train_loss:0.00013, Test_loss:0.00059
learning rate = 0.03455 Epoch:30, Train_loss:0.00012, Test_loss:0.00058
learning rate = 0.03159 Epoch:31, Train_loss:0.00012, Test_loss:0.00057
learning rate = 0.02871 Epoch:32, Train_loss:0.00011, Test_loss:0.00056
learning rate = 0.02591 Epoch:33, Train_loss:0.00011, Test_loss:0.00055
learning rate = 0.02321 Epoch:34, Train_loss:0.00011, Test_loss:0.00055
learning rate = 0.02061 Epoch:35, Train_loss:0.00011, Test_loss:0.00055
learning rate = 0.01813 Epoch:36, Train_loss:0.00012, Test_loss:0.00055
learning rate = 0.01577 Epoch:37, Train_loss:0.00012, Test_loss:0.00055
learning rate = 0.01355 Epoch:38, Train_loss:0.00012, Test_loss:0.00056
learning rate = 0.01147 Epoch:39, Train_loss:0.00012, Test_loss:0.00056
learning rate = 0.00955 Epoch:40, Train_loss:0.00013, Test_loss:0.00057
learning rate = 0.00778 Epoch:41, Train_loss:0.00013, Test_loss:0.00058
learning rate = 0.00618 Epoch:42, Train_loss:0.00014, Test_loss:0.00058
learning rate = 0.00476 Epoch:43, Train_loss:0.00014, Test_loss:0.00059
learning rate = 0.00351 Epoch:44, Train_loss:0.00014, Test_loss:0.00059
learning rate = 0.00245 Epoch:45, Train_loss:0.00014, Test_loss:0.00059
learning rate = 0.00157 Epoch:46, Train_loss:0.00014, Test_loss:0.00060
learning rate = 0.00089 Epoch:47, Train_loss:0.00014, Test_loss:0.00060
learning rate = 0.00039 Epoch:48, Train_loss:0.00014, Test_loss:0.00060
learning rate = 0.00010 Epoch:49, Train_loss:0.00014, Test_loss:0.00060
learning rate = 0.00000 Epoch:50, Train_loss:0.00014, Test_loss:0.00060
==================== Done ====================

四、模型評估

1. Loss圖片

import matplotlib.pyplot as plt
from datetime import datetime
current_time = datetime.now() # 獲取當前時間plt.figure(figsize=(5, 3),dpi=120)plt.plot(train_loss    , label='LSTM Training Loss')
plt.plot(test_loss, label='LSTM Validation Loss')plt.title('Training and Validation Loss')
plt.xlabel(current_time) # 打卡請帶上時間戳，否則代碼截圖無效
plt.legend()
plt.show()

2. 調用模型進行預測

predicted_y_lstm = sc.inverse_transform(model(X_test).detach().numpy().reshape(-1,1))                    # 測試集輸入模型進行預測
y_test_1         = sc.inverse_transform(y_test.reshape(-1,1))
y_test_one       = [i[0] for i in y_test_1]
predicted_y_lstm_one = [i[0] for i in predicted_y_lstm]plt.figure(figsize=(5, 3),dpi=120)
# 畫出真實數據和預測數據的對比曲線
plt.plot(y_test_one[:2000], color='red', label='real_temp')
plt.plot(predicted_y_lstm_one[:2000], color='blue', label='prediction')plt.title('Title')
plt.xlabel('X')
plt.ylabel('Y')
plt.legend()
plt.show()

3. R2值評估

from sklearn import metrics
"""
RMSE ：均方根誤差  ----->  對均方誤差開方
R2   ：決定系數，可以簡單理解為反映模型擬合優度的重要的統計量
"""
RMSE_lstm  = metrics.mean_squared_error(predicted_y_lstm_one, y_test_1)**0.5
R2_lstm    = metrics.r2_score(predicted_y_lstm_one, y_test_1)print('均方根誤差: %.5f' % RMSE_lstm)
print('R2: %.5f' % R2_lstm)

均方根誤差: 6.92733
R2: 0.83259

總結：

本周主要學習了LSTM模型，并且通過實踐更加深入地了解到了LSTM模型。