InitNormal

InitNormal#

class seli.net.InitNormal(init: Literal['Unit', 'He', 'Xavier', 'Glorot', 'Kaiming', 'LeCun'] = 'He', shift: float = 0.0, scale: float = 1.0)[source]#

Bases: Init

Initializes values from a normal distribution. The following automatic scaling methods are supported:

  • Glorot, X., & Bengio, Y. (2010). “Understanding the difficulty of training deep feedforward neural networks.” In Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics (pp. 249-256). JMLR.org.

  • He, K., Zhang, X., Ren, S., & Sun, J. (2015). “Delving deep into rectifiers: Surpassing human-level performance on ImageNet classification.”

  • LeCun, Y., Bottou, L., Orr, G. B., & Müller, K. R. (1998). “Efficient backprop.”

The default is He, the same as in PyTorch.

Methods Summary

__call__(key, shape, dtype)

Call self as a function.

Methods Documentation

__call__(key: Key[Array, ''] | UInt32[Array, '2'], shape: tuple[int, ...], dtype: str | type[Any] | dtype | SupportsDType) → Array[source]#

Call self as a function.

Contents