This my implementation of the triplet loss function using tensorflow 2.3 and keras API. I finetuned a pre-trained inception resnet50 architecture on the triplet loss function to detect faces using considerably less amount of training data as compare to other neural networks