DioscoriDESS

1. Make a directory where your input files (.yaml) and output files (.h5) will be stored:

   mkdir -p /MY_PATH/dioscoridess/cfgs
   mkdir -p /MY_PATH/dioscoridess/logs
   mkdir -p /MY_PATH/dioscoridess/ckpt
   

2. Create your yaml file and save in /MY_PATH/dioscoridess/cfgs/my_input.yaml:

   # Sample ./yaml config file
   common:
   seed: 1234
   vis_GPU: '1' #single GPU compatibility
   log_path: './experiments/training/logs/'
   save_path: './experiments/training/ckpt/'
   print_freq: 25
   
   data_train:
   data_root: '/data/bigbone5/JacobOeding/DL_code/splits/axial_train.csv'
   batch_size: 4
   im_dims: !!python/tuple [512, 512]
   crop : !!python/tuple [0, 0, 0, 0]
   num_classes: 3
   idx_classes : [0,1,2]
   num_channels: 1
   normalization_file: '/data/bigbone5/JacobOeding/DL_code/dataInfo/Axial_MinMaxDataset.csv'
   
   data_val:
   data_root : '/data/bigbone5/JacobOeding/DL_code/splits/axial_val_onlySegmentedSlices.csv'
   batch_size: 4
   im_dims: !!python/tuple [512, 512]
   crop : !!python/tuple [0, 0, 0, 0]
   num_classes: 3
   idx_classes : [0,1,2]
   num_channels: 1
   normalization_file: '/data/bigbone5/JacobOeding/DL_code/dataInfo/Axial_MinMaxDataset.csv'
   evaluate_mode: True
   
   dataLoader_folder: 'dataLoader'
   learn:
   max_steps: 500000 # instead of epochs
   lr : 0.00005
   optimizer: 'adam'
   loss: 'dice_loss_softmax'
   metrics: 'dice_score'
   weights: [1,1.2,1]
   num_classes: 3
   dataloader: 'data_loader'
   comp: ['femur', 'tibia', 'background']
   key_slice: !!python/tuple [0, 0, 0]
   patience: 15 # in number of validation runs
   keep_prob : 0.95
   val_freq : 75 # in steps
   
   model_folder: VNet
   model: 'VNet'
   model_params:
   # github.com/MiguelMonteiro/VNet-Tensorflow
   num_classes: 3
   num_channels : 32
   num_levels : 5
   num_convolutions : !!python/tuple [1, 1, 1, 1, 1]
   bottom_convolutions : 2
   
   pretrain:
   flag: False
   ckpt: ''
   

3. Activate the python environment:

   bash
   export PATH=/netopt/rhel7/versions/python/Anaconda3-5.2.0/bin:/bin:$PATH
   source activate /netopt/share/bin/local/bone/conda_envs/envs/dioscoridess_tf-1.12_py-3.6
   

4. Run DioscoriDESS:

   cd /data/i2/software/dioscoridess
   python main.py --cfg /MY_PATH/dioscoridess/cfgs/my_input.yaml --desc test