Medical_Unet_Dashboard

All the dependencies needs to be installed

pip install requirements.txt

Create an empty folder 'weight'

mkdir weight

Add the dataset folder

Name of the dataset folder should be 'dataset' and the Image folder should be 'image' and Mask folder should be 'mask'

Dataset Folder should look something as shown

Dataset Folder

Overall Working folder structure must be as shown below

Working Directory

Then run the app

streamlit run main.py

Open the network URL generted on any browser

Landing Dashboard

Set the training parameters by selecting number of epochs and learning rate

The training must start automatically after loading data and the logs would be displayed

Training and Validation loss per epoch

Evaluation

Select the train checkpoint you want to visualise the results of and upload image to see it's mask