MODULE_NAME=nnunet_for_pytorch
MODULE_VERSION=21.11.0
WORK_DIR=/proj/nsc_testing/xuan/berzelius-benchmarks/PyTorch/Segmentation/nnUNet
CONTAINER_DIR=/proj/nsc_testing/xuan/containers/${MODULE_NAME}_${MODULE_VERSION}.sif

mkdir -p $WORK_DIR/data $WORK_DIR/results


To download and preprocess the data run:

apptainer exec --nv -B ${WORK_DIR}/data:/data -B ${WORK_DIR}/results:/results --pwd /workspace/nnunet_pyt $CONTAINER_DIR python download.py --task 01  

apptainer exec --nv -B ${WORK_DIR}/data:/data -B ${WORK_DIR}/results:/results --pwd /workspace/nnunet_pyt $CONTAINER_DIR  python /workspace/nnunet_pyt/preprocess.py --task 01 --dim 2
 
Start benchmarking:

apptainer exec --nv -B ${WORK_DIR}/data:/data -B ${WORK_DIR}/results:/results --pwd /workspace/nnunet_pyt $CONTAINER_DIR python scripts/benchmark.py --mode train --gpus 1 --dim 2 --batch_size 256 --amp

apptainer exec --nv -B ${WORK_DIR}/data:/data -B ${WORK_DIR}/results:/results --pwd /workspace/nnunet_pyt $CONTAINER_DIR python scripts/benchmark.py --mode predict --gpus 1 --dim 2 --batch_size 256 --amp

################# Issues #################
# Known issue https://github.com/NVIDIA/DeepLearningExamples/issues/1113
ImportError: cannot import name 'get_num_classes' from 'torchmetrics.utilities.data' (/opt/conda/lib/python3.8/site-packages/torchmetrics/utilities/data.py)

Solution: pip install pytorch-lightning==1.5.10, another error raised when benchmarking predict:
Traceback (most recent call last):
  File "main.py", line 110, in <module>
    trainer.current_epoch = 1
AttributeError: can't set attribute

Solution: pip install torchmetrics==0.6.0, works.

# Muiti-node is not supported in 21.11.0 yet but only in the most recent code on GitHub.

Solution: commenting the L32-33 in the main.py

# 2023-09-28: Not working on the devel reservation