This repository is my practice codes from Object Detection Series from Aladdin Persson.
I made two minor adjustments following the video tutorial.
- In the
loss.py
, it is unnecessary to calltorch.flatten
sinceself.mse
calculates the Mean Squared Error of every entry and adds together that results in a single number. As a result, I did not utilize it. - In the
dataset.py
, I have created thecollate_fn
method to transform the batch of images instead of a single image as I believe that results in a faster process.