CACTI Code Library Documentation

folder configs

• folder base: model and basic dataset
• Each reconstruction algorithm configuration files

folder cacti

• datasets (data preprocessing)
• models (reconstruction algorithms)
• utils (universial function, such as PSNR SSIM Calculation)

folder tools

• train.py (Model Training)
• test_deeplearning.py (Testing end to end deep learning algorithms or deep unfolding algorithms on grayscale or colored simulation dataset)
• test_iterative.py (Testing iterative algorithms or plug and play algorithms on grayscale simulation dataset)
• test_color_iterative.py (Testing iterative algorithms or plug and play algorithms on colored simulation dataset)
• real_data (Testing real data)
• params_flops.py (Statistics of model parameters and FLOPs)
• video_gif (images to video and images to gif transfer)
• onnx_tensorrt (onnx, tensorrt model transfer and testing)

folder test_datasets

• mask (mask for different models)
• simulation (Testing results for 6 benchmark grayscale simulation dataset)
• middle_scale (Testing results for 6 benchmark colored simulation dataset)
• real_data (real data, compress ration from 10 to 50)