tensorplay.nn.init

`tensorplay.nn.init.calculate_gain`	Return the recommended gain value for the given nonlinearity.
`tensorplay.nn.init.uniform_`	Fill `tensor` in-place with samples from U(a, b).
`tensorplay.nn.init.normal_`	Fill `tensor` in-place with samples from N(mean, std^2).
`tensorplay.nn.init.constant_`	Fill `tensor` in-place with the scalar `val`.
`tensorplay.nn.init.ones_`	Fill `tensor` in-place with the scalar value 1.
`tensorplay.nn.init.zeros_`	Fill `tensor` in-place with the scalar value 0.
`tensorplay.nn.init.eye_`	Fill the 2-D `tensor` in-place with the identity matrix.
`tensorplay.nn.init.dirac_`	Fill the {3, 4, 5}-D `tensor` in-place with Dirac delta kernels.
`tensorplay.nn.init.xavier_uniform_`	Fill `tensor` in-place with U(-a, a) where a = gain * sqrt(6 / (fan_in + fan_out)).
`tensorplay.nn.init.xavier_normal_`	Fill `tensor` in-place with N(0, std^2) where std = gain * sqrt(2 / (fan_in + fan_out)).
`tensorplay.nn.init.kaiming_uniform_`	Fill `tensor` in-place with U(-bound, bound) where bound = gain * sqrt(3 / fan).
`tensorplay.nn.init.kaiming_normal_`	Fill `tensor` in-place with N(0, std^2) where std = gain / sqrt(fan).
`tensorplay.nn.init.trunc_normal_`	Fills the input Tensor with values drawn from a truncated normal distribution.
`tensorplay.nn.init.orthogonal_`	Fill `tensor` in-place with a (semi) orthogonal matrix.
`tensorplay.nn.init.sparse_`	Fill the 2-D `tensor` in-place as a sparse matrix.

Help improve this page