https://github.com/1adrianb/unsupervised-face-representation

Implementation for Pre-training strategies and datasets for facial representation learning, ECCV 2022
https://github.com/1adrianb/unsupervised-face-representation

emotion-recognition face-alignment face-recognition face-representation-learning pytorch representation-learning unsupervised-learning

Last synced: 3 days ago
JSON representation

Implementation for Pre-training strategies and datasets for facial representation learning, ECCV 2022

• Host: GitHub
• URL: https://github.com/1adrianb/unsupervised-face-representation
• Owner: 1adrianb
• License: mit
• Created: 2022-07-19T16:16:28.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-02-16T20:16:04.000Z (9 months ago)
• Last Synced: 2024-02-17T20:26:07.576Z (9 months ago)
• Topics: emotion-recognition, face-alignment, face-recognition, face-representation-learning, pytorch, representation-learning, unsupervised-learning
• Language: Python
• Homepage:
• Size: 86.9 KB
• Stars: 59
• Watchers: 4
• Forks: 5
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Pre-training strategies and datasets for facial representation learning

This is the PyTorch implementation for [Facial Representation Learning (FRL) paper](http://www.adrianbulat.com/downloads/ECCV2022/face_representation_learning.pdf):

```

@inproceedings{bulat2022pre,

  title={Pre-training strategies and datasets for facial representation learning},

  author={Bulat, Adrian and Cheng, Shiyang and Yang, Jing and Garbett, Andrew and Sanchez, Enrique and Tzimiropoulos, Georgios},

  journal={ECCV},

  year={2022}

}

```

## Model Zoo

We provide bellow some of the models trained in a self-supervised manner. More models to be added later on.

  

    

      data

      backbone

      url

    

  

  

    

      VGG

      ResNet 50

       model 

    

    

      VGG (1M)

      ResNet 50

       model 

    

  

Code snippet for loading the weights in a torchvision standard resnet:

```python

import torch

from torchvision.models import resnet50

init_weights = torch.load('flr_r50_vgg_face.pth', map_location=torch.device('cpu'))['state_dict']

converted_weights = {k.replace('module.base_net.', ''):v for k, v in init_weights.items()}

model = resnet50(weights=None)

results = model.load_state_dict(converted_weights, strict=False)

# Note: the classifier layer is not loaded (fc.weight and fc.bias)

# similarly,  the projection layers used for pre-training are discarded.

print(results)

```

## Installation

To use the code, clone the repo and install the following packages:

```bash

git clone https://github.com/1adrianb/unsupervised-face-representation

```

### Requirements

* Python >= 3.8

* Numpy

* pytorch: [install instructions](https://pytorch.org/get-started/locally/)

* torchvision: ``conda install torchvision -c pytorch``

* apex: [install instructions](https://github.com/NVIDIA/apex#installation)

* OpenCV: ``pip install opencv-python``

* H5Py: ``conda install h5py``

* tensorboard: ``pip install tensorboard``

* pandas

Note, if you are using pytorch > 1.10 and experience issues with apex, please see [#1282](https://github.com/NVIDIA/apex/pulls/1282). Alternatively you can switch to the native pytorch amp.

## Training

```bash

bash scripts/run.sh

```

Before running the script make sure to set the appropiate paths. The models released in the paper were trained using 64 K40 GPUs.

## Acknowledgement

We thank the original authors for releasing their code: [SwAV](https://github.com/facebookresearch/swav), [MoCo](https://github.com/facebookresearch/moco), [BYOL](https://github.com/deepmind/deepmind-research/blob/master/byol/byol_experiment.py), and [vissl](https://github.com/facebookresearch/vissl) which we base our code base upon.