Update Part1_TensorFlow.ipynb

[marius190]

row_vector = matrix[1]
column_vector = matrix[:,2]
scalar = matrix[1, 2]

You cant acces third dimension (2) in a 2x2 matrix

[alaneuler]

matrix is actually defined by yourself in previous cell, just code it bigger than 2

[nightcap1979]

marius190 is right, the following has to replace the current code:
row_vector = matrix[0]
column_vector = matrix[:,1]
scalar = matrix[0, 1]

[datasith]

Technically, @alaneuler is right. The description uses the term "dimensions" which in this context is interchangeable with "axes" or "rank" (see TensorFlow's documentation). Whether this was intentional or not, a solution could be to define a 2xn matrix with n>=3. I do agree that it would be more appropriate to choose indexes like those @nightcap1979 proposed such that it works for all tensors with 2 axes.