ray.rllib.utils.numpy.lstm#

ray.rllib.utils.numpy.lstm(x, weights: numpy.ndarray, biases: numpy.ndarray | None = None, initial_internal_states: numpy.ndarray | None = None, time_major: bool = False, forget_bias: float = 1.0)[source]#

Calculates LSTM layer output given weights/biases, states, and input.

Parameters:

  • x – The inputs to the LSTM layer including time-rank (0th if time-major, else 1st) and the batch-rank (1st if time-major, else 0th).

  • weights – The weights matrix.

  • biases – The biases vector. All 0s if None.

  • initial_internal_states – The initial internal states to pass into the layer. All 0s if None.

  • time_major – Whether to use time-major or not. Default: False.

  • forget_bias – Gets added to first sigmoid (forget gate) output. Default: 1.0.

Returns:

Tuple consisting of 1) The LSTM layer’s output and 2) Tuple: Last (c-state, h-state).