Paper on arxiv: arxiv

Accepted at IEEE ACCESS Journal

Model training: In the paper, we employ MS1MV2 as the training dataset which can be downloaded from InsightFace (MS1M-ArcFace in DataZoo) Download MS1MV2 dataset from insightface on strictly follow the licence distribution

Unzip the dataset and place it in the data folder

Rename the config/config_xxxxxx.py to config/config.py

1. Train PocketNet with ArcFace loss
   • ./train.sh
2. Train PocketNet with template knowledge distillation
   • ./train_kd.sh
3. Train PocketNet with multi-step template knowledge distillation
   • ./train_kd.sh

All code has been trained and tested using Pytorch 1.7.1

1. download the data from their offical webpages.
2. alternative: The evaluation datasets are available in the training dataset package as bin file
3. Rename the configuration file in config directory based on the evaluation model e.g. rename config_PocketNetM128.py to config.py to evaluate the PocketNetM128
4. set the config.rec to dataset folder e.g. data/faces_emore
5. set the config.val_targets for list of the evaluation dataset
6. download the pretrained model from link the previous table
7. set the config.output to path to pretrained model weights
8. run eval/eval.py
9. the output is test.log contains the evaluation results over all epochs

1. Please apply for permissions from NIST before your usage NIST_Request
2. run eval/IJB/runIJBEval.sh

The code of NAS is available under NAS

• Add pretrained model
• Training configuration
• Add NAS code
• Add evaluation results

If you use any of the provided code in this repository, please cite the following paper:

@article{boutros2021pocketnet,
  author    = {Fadi Boutros and
               Patrick Siebke and
               Marcel Klemt and
               Naser Damer and
               Florian Kirchbuchner and
               Arjan Kuijper},
  title     = {PocketNet: Extreme Lightweight Face Recognition Network Using Neural
               Architecture Search and Multistep Knowledge Distillation},
  journal   = {{IEEE} Access},
  volume    = {10},
  pages     = {46823--46833},
  year      = {2022},
  url       = {https://doi.org/10.1109/ACCESS.2022.3170561},
  doi       = {10.1109/ACCESS.2022.3170561},
}

This project is licensed under the terms of the Attribution-NonCommercial-ShareAlike 4.0 
International (CC BY-NC-SA 4.0) license. 
Copyright (c) 2021 Fraunhofer Institute for Computer Graphics Research IGD Darmstadt