the repo contains two noebooks. in load_split_xray_folders.ipynb a dataset of 4 classes cxr images including Normal ( 2943 images)Pneumonia-Bacterial ( 2701 images) Pneumonia-Viral ( 1490 images) and COVID-19( 1153 images) is downloaded from kaggle and is foldered in diffrent folders. in 4_vgg_dense_mbl_incptin.ipynb the stacked model of pretrained VGG, DenseNet, InceptionNet and MobileNet is defined and fit to the data.
