torch.rand
Returns a tensor, comprising from interval [0, 1) in uniform distribution a set of random numbers are drawn. Tensor is defined by a shape parameter sizes.
tmp = torch.rand(2, 3)
torch.randn
Returns a tensor, comprising a set of random numbers extracted from the standard normal distribution (mean 0 and variance 1, i.e., Gaussian white noise) in the. Tensor is defined by a shape parameter sizes.
tmp = torch.randn(2, 3)
torch.normal
Returns a tensor, comprising a set of random numbers extracted from the specified means the mean and standard deviation std discrete normal distribution.
Standard deviation std is a tensor, comprising a standard normal distribution associated with each element of the output difference.
print(torch.normal(mean=0.5, std=torch.arange(1, 6).float()))
torch.linspace
Returns a 1-dimensional tensor, comprising a step points in the interval start and end uniformly spaced.
Tensor length of the output is determined by the steps.
print(torch.linspace(1, 10, steps=4))