This repository includes two jupyter notebooks. The first one retrains the already pre-trained ResNet-50 using transfer learning in order to classify fruits from the Kaggle 360 Fruits challenge (https://www.kaggle.com/moltean/fruits). The architecture will be adapted in order to compute the class activation maps within the second notebook. A detailed description can be seen on my medium articel for this topic (https://medium.com/@brus.patrick63/class-activation-mapping-using-transfer-learning-of-resnet50-e8ca7cfd657e).