快捷導(dǎo)航

Keras在訓(xùn)練期間可視化訓(xùn)練誤差和測(cè)試誤差實(shí)例

更新時(shí)間：2020年06月16日 10:46:32 作者：bebr

這篇文章主要介紹了Keras在訓(xùn)練期間可視化訓(xùn)練誤差和測(cè)試誤差實(shí)例，具有很好的參考價(jià)值，希望對(duì)大家有所幫助。一起跟隨小編過(guò)來(lái)看看吧

詳細(xì)的解釋?zhuān)x者自行打開(kāi)這個(gè)鏈接查看，我這里只把最重要的說(shuō)下

fit() 方法會(huì)返回一個(gè)訓(xùn)練期間歷史數(shù)據(jù)記錄對(duì)象，包含 training error, training accuracy, validation error, validation accuracy 字段，如下打印

# list all data in history
print(history.history.keys())

完整代碼

# Visualize training history
from keras.models import Sequential
from keras.layers import Dense
import matplotlib.pyplot as plt
import numpy
 
# fix random seed for reproducibility
seed = 7
numpy.random.seed(seed)
# load pima indians dataset
dataset = numpy.loadtxt("pima-indians-diabetes.csv", delimiter=",")
# split into input (X) and output (Y) variables
X = dataset[:,0:8]
Y = dataset[:,8]
# create model
model = Sequential()
model.add(Dense(12, input_dim=8, kernel_initializer='uniform', activation='relu'))
model.add(Dense(8, kernel_initializer='uniform', activation='relu'))
model.add(Dense(1, kernel_initializer='uniform', activation='sigmoid'))
 
# Compile model
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
 
# Fit the model
history = model.fit(X, Y, validation_split=0.33, epochs=150, batch_size=10, verbose=0)
 
# list all data in history
print(history.history.keys())
 
# summarize history for accuracy
plt.plot(history.history['acc'])
plt.plot(history.history['val_acc'])
plt.title('model accuracy')
plt.ylabel('accuracy')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()
 
# summarize history for loss
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('model loss')
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()

補(bǔ)充知識(shí)：訓(xùn)練時(shí)同時(shí)輸出實(shí)時(shí)cost、準(zhǔn)確率圖

首先定義畫(huà)圖函數(shù)：

train_prompt = "Train cost"
cost_ploter = Ploter(train_prompt)
def event_handler_plot(ploter_title, step, cost):
 cost_ploter.append(ploter_title, step, cost)
 cost_ploter.plot()

在訓(xùn)練時(shí)如下方式使用：


EPOCH_NUM = 8
# 開(kāi)始訓(xùn)練
lists = []
step = 0
for epochs in range(EPOCH_NUM):
 # 開(kāi)始訓(xùn)練
 for batch_id, train_data in enumerate(train_reader()):    #遍歷train_reader的迭代器，并為數(shù)據(jù)加上索引batch_id
  train_cost,sult,lab,vgg = exe.run(program=main_program,  #運(yùn)行主程序
        feed=feeder.feed(train_data),    #喂入一個(gè)batch的數(shù)據(jù)
        fetch_list=[avg_cost,predict,label,VGG])   #fetch均方誤差和準(zhǔn)確率
  if step % 10 == 0:    
   event_handler_plot(train_prompt,step,train_cost[0])
  # print(batch_id)
  if batch_id % 10 == 0:         #每100次batch打印一次訓(xùn)練、進(jìn)行一次測(cè)試
   p = [np.sum(pre) for pre in sult]
   l = [np.sum(pre) for pre in lab]
   print(p,l,np.sum(sult),np.sum(lab))
   print('Pass:%d, Batch:%d, Cost:%0.5f' % (epochs, batch_id, train_cost[0]))
  step += 1
 # 保存模型
 if model_save_dir is not None:
  fluid.io.save_inference_model(model_save_dir, ['images'], [predict], exe)

print('訓(xùn)練模型保存完成！')
end = time.time()
print(time.strftime('V100訓(xùn)練用時(shí)：%M分%S秒',time.localtime(end-start)))

實(shí)時(shí)顯示準(zhǔn)確率用同樣的方法

以上這篇Keras在訓(xùn)練期間可視化訓(xùn)練誤差和測(cè)試誤差實(shí)例就是小編分享給大家的全部?jī)?nèi)容了，希望能給大家一個(gè)參考，也希望大家多多支持腳本之家。

您可能感興趣的文章: