The following codes contain an LSTM model with a callback that stops the training once the training set accuracy reached 90%.
It was working fine when I used tensorfl