Semantic Segmentation - UNet 2

ssTem Dataset : dataset/train-volume.tif,train-labels.tif,test-volume.tif

How to Setup and Train

0. clone repository

git clone https://github.com/realJun9u/unet_sstem.git

1. Install Dependancy

# in keras or pytorch
pip install -r requirement.txt

2. Setup Dataset (Seperate Datasets in ./dataset)

python setup.py

3. Train and Evaluate Model

python train.py

4. Predict Model (Create Prediction Images in result directory)

python predict.py

5. Visualize Model

python visualize.py

6. Analyze Result

tensorboard --logdir=logs --host={host} --port={port}

CUDA 버전, cuDNN 버전 확인

1. CUDA 버전

nvcc -V

2. cuDNN 버전

cat /usr/local/cuda/include/cudnn_version.h | grep CUDNN_MAJOR -A 2