QUICKSTART.md

Quickstart instructions

Setting up environment

1. Create conda environment and activate it

   conda create --name GCN_MULTI_env python=3.11
   conda activate GCN_MULTI_env

2. Install requirements

   pip install -r requirements.txt 

Preprocessing CAMUS DATA

1. Download the CAMUS dataset folder from https://humanheart-project.creatis.insa-lyon.fr/database/#collection/6373703d73e9f0047faa1bc8/folder/63fde55f73e9f004868fb7ac
2. Extract the downloaded folder and place the database_nifti folder in data/local_data
3. Run PYTHONPATH=./ python tools/preprocess_CAMUS_displacement.py

Evaluation of trained model:

1. Download the trained model from https://huggingface.co/gillesvdv/GCN_with_displacement_camus_cv1
2. place the downloaded .pth file in experiments/logs/CAMUS_displacement_cv_1/GCN_multi_displacement_small/mobilenet2/trained/
3. Run python eval.py
4. The results will be saved in the folder experiments/logs/CAMUS_displacement_cv_1/GCN_multi_displacement_small/mobilenet2/trained/weights_CAMUS_displacement_cv_1_GCN_multi_displacement_small_best_loss_eval_on_CAMUS_displacement_cv_1/. The plots subfolder contains the resulting plots and predictions.pkl contains the predictions of each sample.

Training your own model:

1. Run python train.py (this will take a long time as the default trains for 5000 epochs)
2. Change the WEIGHTS parameter in files/configs/Eval_CAMUS_displacement.yaml to the path of checkpoint of the trained model in experiments/logs/your_dataset/mobilenetv2/your_run_id/your_weights.pth where your_dataset is the name of the dataset you trained on and your_weights is the name of the checkpoint you want to use, and your_run_id is the automatically generated id of the run you want to use.