Running on Windows within Miniconda3 I get the warning below. Does val_acc need to be changed to val_accuracy in the early stopping rules?
C:\Users\...\miniconda3\envs\dvf\lib\site-packages\keras\callbacks\callbacks.py:846: RuntimeWarning: Early stopping conditioned on metricval_accwhich is not available. Available metrics are: val_loss,val_accuracy,loss,accuracy (self.monitor, ','.join(list(logs.keys()))), RuntimeWarning
Thanks for reporting the issue. Keras has been updated so it is possible that the semantics changed. Hope it worked after you changed it to val_accuracy.
Running on Windows within Miniconda3 I get the warning below. Does val_acc need to be changed to val_accuracy in the early stopping rules?
C:\Users\...\miniconda3\envs\dvf\lib\site-packages\keras\callbacks\callbacks.py:846: RuntimeWarning: Early stopping conditioned on metric
val_accwhich is not available. Available metrics are: val_loss,val_accuracy,loss,accuracy (self.monitor, ','.join(list(logs.keys()))), RuntimeWarning