ray.rllib.utils.numpy.convert_to_numpy#

ray.rllib.utils.numpy.convert_to_numpy(x: numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor | dict | tuple, reduce_type: bool = True) → numpy.array | jnp.ndarray | tf.Tensor | torch.Tensor | dict | tuple[source]#

Converts values in stats to non-Tensor numpy or python types.

Parameters:

  • x – Any (possibly nested) struct, the values in which will be converted and returned as a new struct with all torch/tf tensors being converted to numpy types.

  • reduce_type – Whether to automatically reduce all float64 and int64 data into float32 and int32 data, respectively.

Returns:

A new struct with the same structure as x, but with all values converted to numpy arrays (on CPU).