1.np.random.rand(d0,d1,...)
2.np.random.randn(d0,d1,...)
d0,d1,... refer to the dimension of matrix you want to generate, "rand" generates number belong to [0, 1)(uniform distribution) and "randn" generates number~N(0,1)(normal distribution)
if you don't assign any value to d0,d1,.... then 1 and 2 return a number with "float" type