- python 3.10.8
- torch 2.1.2
- torchvision 0.16.2
- tensorboard 2.15.1
git clone https://github.com/suiyizhao/BaselineIR.git
cd BaselineIR
pip install -r requirements.txt
cd src
Please ensure that the data organization matches the code format for train & test or the code format for infer.
python train.py --data_source /your/dataset/path --experiment your_experiment_name
python test.py --data_source /your/dataset/path --experiment your_experiment_name --model_path /your/model/path --train_crop your_crop_size_in_training --save_image
python infer.py --data_source /your/dataset/path --experiment your_experiment_name --model_path /your/model/path --train_crop your_crop_size_in_training --save_image
# During training, it is recommended to debug first to make sure the code is working properly
python train.py --data_source /your/dataset/path --experiment your_experiment_name --debug
# Manually modify the set_random_seed (utils.py) function by setting "deterministic=True"
set_random_seed(opt.seed, deterministic=True)
python train.py --data_source /your/dataset/path --experiment your_experiment_name --data_parallel
python train.py --data_source /your/dataset/path --experiment your_experiment_name --pretrained /pretrained/model/path
python train.py --data_source /your/dataset/path --experiment your_experiment_name --resume