Make cool Corona Virus detection web app on chest X-Ray images from Pneumonia dataset
1- Training on TPU
2- Transfer Learning
3- Testing three famous CNN architectures (VGG16, InceptionV3 and Xception) 🤓
4- Working with GCS
5- Deploying model as a web app with Anvil
watch this for making web app
1- VGG16 : accuracy: 1.0000 - precision: 1.0000 - recall: 1.0000 - val_loss: 0.0325 - val_accuracy: 0.9873 - val_precision: 0.9972 - val_recall: 0.9849
2- Inception V3 : accuracy: 1.0000 - precision: 1.0000 - recall: 1.0000 - val_loss: 0.0933 - val_accuracy: 0.9795 - val_precision: 0.9876 - val_recall: 0.9835
3- Xcpetion : accuracy: 1.0000 - precision: 1.0000 - recall: 1.0000 - val_loss: 0.0620 - val_accuracy: 0.9785 - val_precision: 0.9917 - val_recall: 0.9782
Colab notebook is provided, simply run the cells to re produce the results
For the Web App simply create a quick app on anvil (tutorial link provided above). Client side code is in the colab notebook. Server code is in server file
Pretrained VGG16 model weights