Kaiming¶

class olympus.models.inits.kaiming.Kaiming[source]¶

Bases: olympus.models.inits.base.Initialization

References

[1]	Xavier Glorot, Yoshua Bengio, “Understanding the difficulty of training deep feedforward neural networks”

[2]	Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun “Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification”, Feb 2015

Methods

__call__(model) Init model using given function for Linear and Conv2d, and {0, 1} for BatchNorm.

layer_init

class olympus.models.inits.kaiming.KaimingNormal(a, mode, non_linearity)[source]¶

Bases: olympus.models.inits.base.Initialization

See :class`.Kaiming`

Methods

__call__(model) Call self as a function.

class olympus.models.inits.kaiming.KaimingUniform(a, mode, non_linearity)[source]¶

See :class`.Kaiming`

Methods

__call__(model) Init model using given function for Linear and Conv2d, and {0, 1} for BatchNorm.