我在使用tf.callbacks.ModelCheckpoint时遇到了问题。如您在我的日志文件中所见，警告总是出现在最后一次迭代之前，而val_acc是在最后一次迭代中计算的。因此，ModelCheckpoint永远找不到val_acc。
Epoch 1/301/8 [==>...........................] - ETA: 19s - loss: 1.4174 - accuracy: 0.30002/8 [======>.......................] - ETA: 8s - loss: 1.3363 - accuracy: 0.3500 3/8 [==========>...................] - ETA: 4s - loss: 1.3994 - accuracy: 0.26674/8 [==============>...............] - ETA: 3s - loss: 1.3527 - accuracy: 0.32506/8 [=====================>........] - ETA: 1s - loss: 1.3042 - accuracy: 0.3333WARNING:tensorflow:Can save best model only with val_acc available, skipping.8/8 [==============================] - 4s 482ms/step - loss: 1.2846 - accuracy: 0.3375 - val_loss: 1.3512 - val_accuracy: 0.5000Epoch 2/301/8 [==>...........................] - ETA: 0s - loss: 1.0098 - accuracy: 0.50003/8 [==========>...................] - ETA: 0s - loss: 0.8916 - accuracy: 0.53335/8 [=================>............] - ETA: 0s - loss: 0.9533 - accuracy: 0.56006/8 [=====================>........] - ETA: 0s - loss: 0.9523 - accuracy: 0.56677/8 [=========================>....] - ETA: 0s - loss: 0.9377 - accuracy: 0.5714WARNING:tensorflow:Can save best model only with val_acc available, skipping.8/8 [==============================] - 1s 98ms/step - loss: 0.9229 - accuracy: 0.5750 - val_loss: 1.2507 - val_accuracy: 0.5000这是我训练CNN的代码。
callbacks = [ TensorBoard(log_dir=r'C:\Users\reda\Desktop\logs\{}'.format(Name), histogram_freq=1), ModelCheckpoint(filepath=r"C:\Users\reda\Desktop\checkpoints\{}".format(Name), monitor='val_acc', verbose=2, save_best_only=True, mode='max')]history = model.fit_generator( train_data_gen, steps_per_epoch=total_train // batch_size, epochs=epochs, validation_data=val_data_gen, validation_steps=total_val // batch_size, callbacks=callbacks)回答：
我知道这些事情有时会让人感到沮丧…但TensorFlow要求您明确写出您想要计算的度量指标的名称。
您需要明确指出’val_accuracy’。
metric = 'val_accuracy'ModelCheckpoint(filepath=r"C:\Users\reda.elhail\Desktop\checkpoints\{}".format(Name), monitor=metric, verbose=2, save_best_only=True, mode='max')]希望这对您有帮助 =)
*** 正如BlueTurtle后来指出的（请给他们的回答点赞，他们的回答可能还在下面），您还需要使用完整的度量指标名称，以匹配您的model.compile、ModelCheckpoint和EarlyStopping。
