torch.nn.init.xavier_normal_
function xavier_normal_(tensor: Tensor, options?: XavierOptions): Tensorfunction xavier_normal_(tensor: Tensor, gain: number, options?: XavierOptions): TensorFill the input Tensor with values using a Xavier normal distribution.
The method is described in "Understanding the difficulty of training deep feedforward neural networks" - Glorot, X. & Bengio, Y. (2010).
Values are sampled from N(0, std^2) where std = gain * sqrt(2 / (fan_in + fan_out))
Also known as Glorot initialization.
Parameters
tensorTensor- An n-dimensional Tensor
optionsXavierOptionsoptional- Optional settings for Xavier initialization
Returns
Tensor– The input tensor with Xavier normal initializationExamples
const w = torch.empty(3, 5);
torch.nn.init.xavier_normal_(w, { gain: 1.0 });