clarena.cl_algorithms.regularisers
Continual Learning Regularisers
This submodule provides the regularisers which are added to the loss function of corresponding continual learning algorithms in CLArena. It can promote forgetting preventing which is the major mechanism in regularisation-based approaches, or for other purposes.
Please note that this is an API documantation. Please refer to the main documentation pages for more information about the regularisers:
- Implement your regularisers in CL algorithms
- A Beginners' Guide to Continual Learning (Regularisation-based Approaches)
The regularisers are implemented as subclasses of nn.Module.
1r""" 2 3# Continual Learning Regularisers 4 5This submodule provides the **regularisers** which are added to the loss function of corresponding continual learning algorithms in CLArena. It can promote forgetting preventing which is the major mechanism in regularisation-based approaches, or for other purposes. 6 7Please note that this is an API documantation. Please refer to the main documentation pages for more information about the regularisers: 8 9- [**Implement your regularisers in CL algorithms**](https://pengxiang-wang.com/projects/continual-learning-arena/docs/implement-your-cl-modules/cl-algorithm#sec-regularisers) 10- [**A Beginners' Guide to Continual Learning (Regularisation-based Approaches)**](https://pengxiang-wang.com/posts/continual-learning-beginners-guide#sec-regularisation-based-approaches) 11 12 13The regularisers are implemented as subclasses of `nn.Module`. 14 15""" 16 17from .distillation import DistillationReg 18from .hat_mask_sparsity import HATMaskSparsityReg 19from .parameter_change import ParameterChangeReg 20 21__all__ = ["distillation", "hat_mask_sparsity", "parameter_change"]