Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/facebookresearch/fair_self_supervision_benchmark
Scaling and Benchmarking Self-Supervised Visual Representation Learning
https://github.com/facebookresearch/fair_self_supervision_benchmark
Last synced: about 2 months ago
JSON representation
Scaling and Benchmarking Self-Supervised Visual Representation Learning
- Host: GitHub
- URL: https://github.com/facebookresearch/fair_self_supervision_benchmark
- Owner: facebookresearch
- License: other
- Archived: true
- Created: 2019-05-13T21:38:48.000Z (over 5 years ago)
- Default Branch: main
- Last Pushed: 2021-10-12T21:34:03.000Z (almost 3 years ago)
- Last Synced: 2024-07-05T16:28:13.979Z (3 months ago)
- Language: Python
- Size: 49.8 MB
- Stars: 587
- Watchers: 23
- Forks: 63
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
Awesome Lists containing this project
- awesome-self-supervised-learning - [repo - supervision approaches. (Computer Vision / Image Representation Learning)
README
## FAIR Self-Supervision Benchmark is deprecated. Please see [VISSL](https://vissl.ai/), a ground-up rewrite of benchmark in [PyTorch](https://pytorch.org/).
# FAIR Self-Supervision Benchmark
This code provides various benchmark (and legacy) tasks for evaluating quality
of visual representations learned by various self-supervision approaches. This code corresponds to our work on [Scaling and Benchmarking Self-Supervised Visual Representation Learning](https://arxiv.org/abs/1905.01235). The code is written in Python and can be used to **evaluate both PyTorch and Caffe2 models** (see [this](https://github.com/facebookresearch/fair_self_supervision_benchmark/tree/master/extra_scripts#using-pytorch-models)). We hope that this
benchmark release will provided a consistent evaluation strategy that will allow
measuring the progress in self-supervision easily.
## Introduction
The goal of fair_self_supervision_benchmark is to standardize the methodology for evaluating quality of visual representations learned by various self-supervision approaches. Further, it provides evaluation on a variety of tasks as follows:
**Benchmark tasks**: The benchmark tasks are based on principle: a good representation (1) transfers to *many* different tasks, and, (2) transfers with *limited* supervision and *limited* fine-tuning. The tasks are as follows.
- Image Classification
- [VOC07](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf)
- [COCO2014](https://arxiv.org/abs/1405.0312)
- [Places205](http://places.csail.mit.edu/places_NIPS14.pdf)
- Low-Shot Image Classification
- [VOC07](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf)
- [Places205](http://places.csail.mit.edu/places_NIPS14.pdf)
- Object Detection on [VOC07](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf) and [VOC07+12](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf) with frozen backbone for detectors:
- [Fast R-CNN](https://arxiv.org/abs/1504.08083)
- [Faster R-CNN](https://arxiv.org/abs/1506.01497)
- [Surface Normal Estimation](https://web.eecs.umich.edu/~fouhey/2013/3dp/index.html)
- Visual Navigation in [Gibson](https://arxiv.org/abs/1808.10654) Environment
These Benchmark tasks use the network architectures:
- [AlexNet](https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf)
- [ResNet50](https://arxiv.org/abs/1512.03385)
**Legacy tasks:** We also classify some commonly used evaluation tasks as legacy tasks for reasons mentioned in Section 7 of [paper](https://arxiv.org/abs/1905.01235). The tasks are as follows:
- [ImageNet-1K](http://www.image-net.org/papers/imagenet_cvpr09.pdf) classification task
- [VOC07](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf) full finetuning
- Object Detection on [VOC07](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf) and [VOC07+12](http://host.robots.ox.ac.uk/pascal/VOC/pubs/everingham10.pdf) with full tuning for detectors:
- [Fast R-CNN](https://arxiv.org/abs/1504.08083)
- [Faster R-CNN](https://arxiv.org/abs/1506.01497)
## License
fair_self_supervision_benchmark is CC-NC 4.0 International licensed, as found in the LICENSE file.
## Citation
If you use fair_self_supervision_benchmark in your research or wish to refer to the baseline results published in the [paper](https://arxiv.org/abs/1905.01235), please use the following BibTeX entry.
```
@article{goyal2019scaling,
title={Scaling and Benchmarking Self-Supervised Visual Representation Learning},
author={Goyal, Priya and Mahajan, Dhruv and Gupta, Abhinav and Misra, Ishan},
journal={arXiv preprint arXiv:1905.01235},
year={2019}
}
```
## Installation
Please find installation instructions in [`INSTALL.md`](INSTALL.md).
## Getting Started
After installation, please see [`GETTING_STARTED.md`](GETTING_STARTED.md) for how to run various benchmark tasks.
## Model Zoo
We provide models used in our [paper](https://arxiv.org/abs/1905.01235) in the [`MODEL_ZOO`](MODEL_ZOO.md).
## References
- [Scaling and Benchmarking Self-Supervised Visual Representation Learning](https://arxiv.org/abs/1905.01235). Priya Goyal, Dhruv Mahajan, Abhinav Gupta*, Ishan Misra*. Tech report, arXiv, May 2019.