27/07/2023
(Un)Fortunately, I don't need to do resizing as part of preprocessing. It is done already.
What did I do/learn till now:
- Got it running on the GPU. On Windows. What a mess. 2. I don't want to dual boot. As much as it would make things easy.
- Set basic params like image and batch size which will be needed for training
- A subset (20%) of training will be used for validation
- Got the detection in the training folders right
# ['hotdog', 'nothotdog']
- Visualized the first 9 images in the training dataset