How do I assign a 3d numpy array of different sizes to each element of a numpy vector?

Question

Here are the situations:

x.shape = (20,)

x1 = x [0,]
x2 = x [1,]
x3 = x [2,]

...

The shapes of x1, x2, x3 ... are all different.

For example,

x1.shape = (300, 400, 3)
x2.shape = (280, 520, 3)
x3.shape = (330, 400, 3)

I have a 3d numpy array of different sizes

I want to assign it to each element of vector called x.

What should I do?

Answer 1

Make your elements first, then create your x array from them

x = numpy.array([x1, x2, x3])

A modern numpy will automatically use dtype=object whereas in previous versions you had to add dtype=object explicitly.

This isn't really much different from using a plain Python list though:

x = [x1, x2, x3]