Based on the original work by @CrimyTheBold and their post on Medium:
We are reimplementing this for a 1D dataset as opposed to the MNIST numbers it was originally designed for.
- 02 tripletloss MNIST : building a similarity comparer using triplet loss function on MNIST.