Final project for DIP 2024 Lectures

Table of contents

1. Overview
2. Technical Report
3. Installation
4. Metric
5. Training
6. Inference
7. Logs and Visualization

• Prerequisites : Before take a deep dive into the project, ensure that you'd installed all requirements packages by this command : If you have CUDA kernel acceleration, install Pytorch with CUDA

  • I recommend the version of Pytorch and CUDA is 1.13 and 11.6, check the Pytorch Homepage for more details !

  • After that , run this command to install dependecies :

    pip install -r requirements.txt
    

• There are two type of model wrapper architecture : SegTransVAE and MONAI wrapper model

  • For SegTransVAE training , move the work directory to .... and run this command :

      python/python3 train.py
    

  • For MONAI wrapper model training , move the work directory to .... and run this command :

      python/python3 train.py --model_type ... --batch_size ... --epochs ... --volume ... --test_size ...
    

    Where the model_type can be : UNet, Segresnet, SWINUNETR, UNETR and so on. You can modify the code for more choice of model type

---

license: apache-2.0