This assignment will be closed on May 25, 2026 (23:59:59).
You must be authenticated to submit your files
CSC_43042_EP - TD9: Neural networks from scratch
Download and unzip the archive CSC43042EP-td9-1-handin.zip.
To achieve full grade, you must complete parts 1 and 2.1 (up to and
including python grader.py 4).
1. Build an un-trainable network: feed-forward
Activations: Upload your solutions to 1.1 to 1.3
Test your work with python grader.py 1 and upload it
using the box below:
Upload form is only available when connected
1.4 Feed-forward
Upload your solution to 1.4.1 Network initialization
Test your work with python grader.py 2 and upload it
using the box below:
Upload form is only available when connected
Upload your solution to 1.4.2 Feed-forward pass
Test your work with python grader.py 3 and upload it
using the box below:
Upload form is only available when connected
2. Actually learning something: back-propagation
Upload your solution to 2.1 Losses
Test your work with python grader.py 4 and upload it
using the box below:
Upload form is only available when connected
2.2 Back-propagation
Upload your solution to 2.2.1 Back-propagation pass to compute gradients
Test your work with python grader.py 5 and upload it
using the box below:
Upload form is only available when connected
Upload your solution to 2.2.2 Update parameters
Test your work with python grader.py 6 and upload it
using the box below:
Upload form is only available when connected
3. Xaggle competition
Upload form is only available when connected