How can I create a tensor of batch batch_size of uniformly distributed values between -1 and 1?

Question

The title pretty much sums it: How can I create a tensor of batch batch_size of uniformly distributed values between -1 and 1 with pytorch?

def create_latent_