README

This is the official Pytorch implementation of "Operating critical machine learning models in resource constrained regimes", Raghavendra Selvan et al. 2023

What is this repository for?

• Reproduce results on LIDC dataset reported in the paper
• v1.0

How do I get set up?

• Basic Pytorch dependency
• Tested on Pytorch 1.7, Python 3.8
• The preprocessed LIDC data is provided as an archive here. Unzip and point that as the data_dir.
• Fine tune a pretrained Densenet model from with half precision using 8-bit optimizer:

python train.py --batch_size 32 --timm --num_epochs 50 --lr 1e-5 --seed 1 --lidc --model_name densenet --bnb --half

Usage guidelines

• Kindly cite our publication if you use any part of the code

@inproceedings{raghav2023Operating,
 	title={Operating critical machine learning models in resource constrained regimes},
	author={Raghavendra Selvan and Julian Schön and Erik B. Dam},
	booktitle={Workshop on Resource Aware Medical Imaging at MICCAI},
	month={July},
 	note={arXiv preprint arXiv:2302.06900},
	year={2023}}

Who do I talk to?

• raghav@di.ku.dk