purepisces
commented
2 months ago In the Detailed Broadcasting Rules section, the "Can't Broadcasting" example is exactly the same as the example above?
Thank you so much for taking the time to carefully read my machine learning blog. I'm thrilled to know that people are engaging with it. I'll continue to improve it, and if you have any other questions, I'm always happy to discuss them!
In the Detailed Broadcasting Rules section, the "Can't Broadcasting" example is exactly the same as the example above?
Yes, Thank you so much for pointing that out! I've modified it to a new version. You can check my progress.txt, which contains all the markdown files that I've already finished. I haven't thoroughly checked the other markdown files yet; I've just added content to them for now.
Example2: Can't Broadcasting
import numpy as np
c = np.array([1, 2, 3])
print(c.shape) #(3,)
new_shape = (3,2)
print(np.broadcast_to(c, new_shape))-
Original Shape:
(3,)- This shape has only one dimension:
3. For broadcasting purposes, it can be treated as(1, 3).
- This shape has only one dimension:
- Target Shape:
(3, 2)
Comparison
- Trailing Dimensions:
3(original) vs.2(target) – these are not compatible because they are not equal and neither is 1. - Leading Dimensions:
1(original) vs.3(target) – these are compatible because 1 can be expanded to 3.
Since the trailing dimensions do not match and are not compatible, broadcasting cannot proceed.
Infinity4B
https://github.com/purepisces/Wenqing-Machine_Learning_Blog/blob/ba4bcf6fe7fafc3aaf67f319d645039ae3781159/Foundational-Concepts-in-Machine-Learning/Linear-Algebra/Numpy-Tips.md?plain=1#L708 In the Detailed Broadcasting Rules section, the "Can't Broadcasting" example is exactly the same as the example above?