Week 2 - Improving Framework Functionality

After struggling to install the onnxsim, I figured out the issue was to do with python3.12-dev not being installed onto WSL2. Since cmake requires Python development files and they didn’t exist on my machine, an error was being thrown. After installing pythn3.12-dev, the project now runs properly. I then updated the documentation detailing my issue and how I fixed it.
Since the framework was supposed to on laptops with GPU and my laptop doesn’t have one, I’ve had to add some extra functionality so I can run the project locally. After making these changes, I opened a pull request so future users are able to utilise the framework.
Added new functionality so intermediate directories are built when saving model performance metrics.
Continued reading and using other materials to understand CNNs and how to optimise our model.

Learning PyTorch syntax to understand how the models in papers I’m reading are implemented.
Trying to change hardcoded values in the project’s code without breaking something.

Figure out a way to safely deal with hardcoded values in the project’s code.
Split the image dataset we use into separate train, test, and validation datasets so we can better understand the model’s effectiveness.