Tensorflow 理解模型训练日志语句_Tensorflow_Machine Learning_Conv Neural Network_Python 3.7

Tensorflow 理解模型训练日志语句

tensorflow machine-learning

Tensorflow 理解模型训练日志语句,tensorflow,machine-learning,conv-neural-network,python-3.7,Tensorflow,Machine Learning,Conv Neural Network,Python 3.7,当我的模特接受训练时，我得到以下信息。我想知道29/30和30/30代表什么 Epoch 1/80 29/30 [============================>.] - ETA: 0s - loss: 1.1177 - acc: 0.5538Epoch 1/80 30/30 [==============================] - 31s 1s/step - loss: 1.1189 - acc: 0.5540 - val_loss: 1.1

当我的模特接受训练时，我得到以下信息。我想知道29/30和30/30代表什么

    Epoch 1/80
    29/30 [============================>.] - ETA: 0s - loss: 1.1177 - acc: 0.5538Epoch 1/80
    30/30 [==============================] - 31s 1s/step - loss: 1.1189 - acc: 0.5540 - val_loss: 1.1182 - val_acc: 0.5741
    Epoch 2/80
    29/30 [============================>.] - ETA: 0s - loss: 1.0194 - acc: 0.6063Epoch 1/80
    30/30 [==============================] - 24s 815ms/step - loss: 1.0149 - acc: 0.6075 - val_loss: 1.0946 - val_acc: 0.6136

首先，这里的29/30、30/30都显示了您的训练进度（这里是fit call）

30/30还显示了一个历元的最后一批，如果您指定val_数据，它还将根据val数据验证模型，这就是为什么30/30中有val_损失和val_acc，但29/30中没有此类信息

Epoch 1/80
29/30 [============================>.] - ETA: 0s - loss: 1.1177 - acc: 0.5538Epoch 1/80
30/30 [==============================] - 31s 1s/step - loss: 1.1189 - acc: 0.5540 - val_loss: 1.1182 - val_acc: 0.5741

29/30是分批进行的进度，29批共完成30次