TensorDictBase.to_padded_tensor#
- TensorDictBase.to_padded_tensor(padding=0.0, mask_key: NestedKey | None = None) Self#
Converts all nested tensors to a padded version and adapts the batch-size accordingly.
- Parameters:
padding (float) – the padding value for the tensors in the tensordict. Defaults to
0.0.mask_key (NestedKey, optional) – if provided, the key where a mask for valid values will be written. Will result in an error if the heterogeneous dimension isn’t part of the tensordict batch-size. Defaults to
None
Every tensordict class and every tensorclass has this method. For the other methods, see Methods by task.