Naive Keras is used to build CNN and Transfer learning is used to create this model which simply means that a pretrained network is used and then layers are added to it. For training,Adam optimizer and the Categorical Cross Entropy loss function is used. The model achieves about 85% accuracy with HAM dataset. Download the datset(Two images folder of approx 1.4GB each names-HAM10000_images_part_1.zip and HAM10000_images_part_2.zip) from - https://www.kaggle.com/kmader/skin-cancer-mnist-ham10000