OmniSafe Normalizer#

Normalizer(shape[, clip])

Calculate normalized raw_data from running mean and std

Normalizer#

Documentation

class omnisafe.common.normalizer.Normalizer(shape, clip=1e6)[source]#

Calculate normalized raw_data from running mean and std

References

Title: Updating Formulae and a Pairwise Algorithm for Computing Sample Variances
Author: Tony F. Chan, Gene H. Golub, Randall J. LeVeque
URL: http://i.stanford.edu/pub/cstr/reports/cs/tr/79/773/CS-TR-79-773.pdf

Initialize the normalize.

__init__(shape, clip=1e6)[source]#: Initialize the normalize.

_push(raw_data)[source]#

Update the mean and std by the raw_data.

Parameters:: raw_data (Tensor) – raw data to be normalized.
Return type:: None

forward(data)[source]#

Normalize the data.

Return type:: Tensor

load_state_dict(state_dict, strict=True)[source]#

Copies parameters and buffers from state_dict into this module and its descendants. If strict is True, then the keys of state_dict must exactly match the keys returned by this module’s state_dict() function.

Parameters:

state_dict (dict) – a dict containing parameters and persistent buffers.
strict (bool, optional) – whether to strictly enforce that the keys in state_dict match the keys returned by this module’s state_dict() function. Default: True

Returns:

``NamedTuple`` with ``missing_keys`` and ``unexpected_keys`` fields –

missing_keys is a list of str containing the missing keys
unexpected_keys is a list of str containing the unexpected keys

Note

If a parameter or buffer is registered as None and its corresponding key exists in state_dict, load_state_dict() will raise a RuntimeError.

property mean: Tensor#: Return the mean of the normalize.

normalize(data)[source]#

Normalize the data.

Hint

If the data is the first data, the data will be used to initialize the mean and std.
If the data is not the first data, the data will be normalized by the mean and std.
Update the mean and std by the data.

Parameters:: data (Tensor) – raw data to be normalized.
Return type:: Tensor

property shape: Tuple[int, ...]#: Return the shape of the normalize.

property std: Tensor#: Return the std of the normalize.