Lstm in kears and tensorflow

what is the Alternative of

1)num_layers ==
2) bidirection=True
in lstm_tensorflow as in lstm_pytorch

if we want to include many layers and want backprpgation in lstm how to implement it in keras or tensorflow