random_uniform initialization for backends

[mganahl]

[summer-bebop]

Hello,

I’d like to help on this one. If I understand correctly (considering the amount of information provided :) ), the goal is to add a method rand_uniform to each backend classes + test + generic backend ?

[mganahl]

that's right, all backend (BaseBackend, NumPyBackend, TensorFlowBackend, PyTorchBackend, JaxBackend, ShellBackend) should implement it (or raise a NotImplementedError)!
The implementation should largely parallel randn. Thanks for your help!!