Tensor.kaiming_normal()
Samples tensor elements according to the method described in He et al., 2015 (opens in a new tab) using a normal distribution.
Usage
from tinygrad.tensor import Tensor
tensor = Tensor.kaiming_normal(3, 3)
print(tensor.numpy())
Return value
[[-0.43342337 0.04244912 0.5609434 ]
[-0.99966764 -0.00925403 1.2957743 ]
[-0.36708504 0.8206572 -1.6821432 ]]