# This is my (2, 3, 4) shape np array
array([[[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]],

       [[12, 13, 14, 15],
        [16, 17, 18, 19],
        [20, 21, 22, 23]]])

# I want to reshape the elements 8-to-15 in new 2D array
# I will convert it to 1D array
d321d = d.flatten()

#This is how it looks now
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
       17, 18, 19, 20, 21, 22, 23])

#Now I will grab the elements that I want
d1 = d321d[8:16]

#This is how it looks now
array([ 8,  9, 10, 11, 12, 13, 14, 15])

# Finally I'm able to reshape above values in 2D array
fd2 = np.reshape(d1, (-1, 4))
#This is how it looks now
#RESULT :
array([[ 8,  9, 10, 11],
       [12, 13, 14, 15]])