PyTorch implementation of SwAV (Swapping Assignments between Views)

Related tags

Deep Learningswav
Overview

Unsupervised Learning of Visual Features by Contrasting Cluster Assignments

This code provides a PyTorch implementation and pretrained models for SwAV (Swapping Assignments between Views), as described in the paper Unsupervised Learning of Visual Features by Contrasting Cluster Assignments.

SwAV Illustration

SwAV is an efficient and simple method for pre-training convnets without using annotations. Similarly to contrastive approaches, SwAV learns representations by comparing transformations of an image, but unlike contrastive methods, it does not require to compute feature pairwise comparisons. It makes our framework more efficient since it does not require a large memory bank or an auxiliary momentum network. Specifically, our method simultaneously clusters the data while enforcing consistency between cluster assignments produced for different augmentations (or “views”) of the same image, instead of comparing features directly. Simply put, we use a “swapped” prediction mechanism where we predict the cluster assignment of a view from the representation of another view. Our method can be trained with large and small batches and can scale to unlimited amounts of data.

Model Zoo

We release several models pre-trained with SwAV with the hope that other researchers might also benefit by replacing the ImageNet supervised network with SwAV backbone. To load our best SwAV pre-trained ResNet-50 model, simply do:

import torch
model = torch.hub.load('facebookresearch/swav:main', 'resnet50')

We provide several baseline SwAV pre-trained models with ResNet-50 architecture in torchvision format. We also provide models pre-trained with DeepCluster-v2 and SeLa-v2 obtained by applying improvements from the self-supervised community to DeepCluster and SeLa (see details in the appendix of our paper).

method epochs batch-size multi-crop ImageNet top-1 acc. url args
SwAV 800 4096 2x224 + 6x96 75.3 model script
SwAV 400 4096 2x224 + 6x96 74.6 model script
SwAV 200 4096 2x224 + 6x96 73.9 model script
SwAV 100 4096 2x224 + 6x96 72.1 model script
SwAV 200 256 2x224 + 6x96 72.7 model script
SwAV 400 256 2x224 + 6x96 74.3 model script
SwAV 400 4096 2x224 70.1 model script
DeepCluster-v2 800 4096 2x224 + 6x96 75.2 model script
DeepCluster-v2 400 4096 2x160 + 4x96 74.3 model script
DeepCluster-v2 400 4096 2x224 70.2 model script
SeLa-v2 400 4096 2x160 + 4x96 71.8 model -
SeLa-v2 400 4096 2x224 67.2 model -

Larger architectures

We provide SwAV models with ResNet-50 networks where we multiply the width by a factor ×2, ×4, and ×5. To load the corresponding backbone you can use:

import torch
rn50w2 = torch.hub.load('facebookresearch/swav:main', 'resnet50w2')
rn50w4 = torch.hub.load('facebookresearch/swav:main', 'resnet50w4')
rn50w5 = torch.hub.load('facebookresearch/swav:main', 'resnet50w5')
network parameters epochs ImageNet top-1 acc. url args
RN50-w2 94M 400 77.3 model script
RN50-w4 375M 400 77.9 model script
RN50-w5 586M 400 78.5 model -

Running times

We provide the running times for some of our runs:

method batch-size multi-crop scripts time per epoch
SwAV 4096 2x224 + 6x96 * * * * 3min40s
SwAV 256 2x224 + 6x96 * * 52min10s
DeepCluster-v2 4096 2x160 + 4x96 * 3min13s

Running SwAV unsupervised training

Requirements

Singlenode training

SwAV is very simple to implement and experiment with. Our implementation consists in a main_swav.py file from which are imported the dataset definition src/multicropdataset.py, the model architecture src/resnet50.py and some miscellaneous training utilities src/utils.py.

For example, to train SwAV baseline on a single node with 8 gpus for 400 epochs, run:

python -m torch.distributed.launch --nproc_per_node=8 main_swav.py \
--data_path /path/to/imagenet/train \
--epochs 400 \
--base_lr 0.6 \
--final_lr 0.0006 \
--warmup_epochs 0 \
--batch_size 32 \
--size_crops 224 96 \
--nmb_crops 2 6 \
--min_scale_crops 0.14 0.05 \
--max_scale_crops 1. 0.14 \
--use_fp16 true \
--freeze_prototypes_niters 5005 \
--queue_length 3840 \
--epoch_queue_starts 15

Multinode training

Distributed training is available via Slurm. We provide several SBATCH scripts to reproduce our SwAV models. For example, to train SwAV on 8 nodes and 64 GPUs with a batch size of 4096 for 800 epochs run:

sbatch ./scripts/swav_800ep_pretrain.sh

Note that you might need to remove the copyright header from the sbatch file to launch it.

Set up dist_url parameter: We refer the user to pytorch distributed documentation (env or file or tcp) for setting the distributed initialization method (parameter dist_url) correctly. In the provided sbatch files, we use the tcp init method (see * for example).

Evaluating models

Evaluate models: Linear classification on ImageNet

To train a supervised linear classifier on frozen features/weights on a single node with 8 gpus, run:

python -m torch.distributed.launch --nproc_per_node=8 eval_linear.py \
--data_path /path/to/imagenet \
--pretrained /path/to/checkpoints/swav_800ep_pretrain.pth.tar

The resulting linear classifier can be downloaded here.

Evaluate models: Semi-supervised learning on ImageNet

To reproduce our results and fine-tune a network with 1% or 10% of ImageNet labels on a single node with 8 gpus, run:

  • 10% labels
python -m torch.distributed.launch --nproc_per_node=8 eval_semisup.py \
--data_path /path/to/imagenet \
--pretrained /path/to/checkpoints/swav_800ep_pretrain.pth.tar \
--labels_perc "10" \
--lr 0.01 \
--lr_last_layer 0.2
  • 1% labels
python -m torch.distributed.launch --nproc_per_node=8 eval_semisup.py \
--data_path /path/to/imagenet \
--pretrained /path/to/checkpoints/swav_800ep_pretrain.pth.tar \
--labels_perc "1" \
--lr 0.02 \
--lr_last_layer 5

Evaluate models: Transferring to Detection with DETR

DETR is a recent object detection framework that reaches competitive performance with Faster R-CNN while being conceptually simpler and trainable end-to-end. We evaluate our SwAV ResNet-50 backbone on object detection on COCO dataset using DETR framework with full fine-tuning. Here are the instructions for reproducing our experiments:

  1. Install detr and prepare COCO dataset following these instructions.

  2. Apply the changes highlighted in this gist to detr backbone file in order to load SwAV backbone instead of ImageNet supervised weights.

  3. Launch training from detr repository with run_with_submitit.py.

python run_with_submitit.py --batch_size 4 --nodes 2 --lr_backbone 5e-5

Common Issues

For help or issues using SwAV, please submit a GitHub issue.

The loss does not decrease and is stuck at ln(nmb_prototypes) (8.006 for 3000 prototypes).

It sometimes happens that the system collapses at the beginning and does not manage to converge. We have found the following empirical workarounds to improve convergence and avoid collapsing at the beginning:

  • use a lower epsilon value (--epsilon 0.03 instead of the default 0.05)
  • carefully tune the hyper-parameters
  • freeze the prototypes during first iterations (freeze_prototypes_niters argument)
  • switch to hard assignment
  • remove batch-normalization layer from the projection head
  • reduce the difficulty of the problem (less crops or softer data augmentation)

We now analyze the collapsing problem: it happens when all examples are mapped to the same unique representation. In other words, the convnet always has the same output regardless of its input, it is a constant function. All examples gets the same cluster assignment because they are identical, and the only valid assignment that satisfy the equipartition constraint in this case is the uniform assignment (1/K where K is the number of prototypes). In turn, this uniform assignment is trivial to predict since it is the same for all examples. Reducing epsilon parameter (see Eq(3) of our paper) encourages the assignments Q to be sharper (i.e. less uniform), which strongly helps avoiding collapse. However, using a too low value for epsilon may lead to numerical instability.

Training gets unstable when using the queue.

The queue is composed of feature representations from the previous batches. These lines discard the oldest feature representations from the queue and save the newest one (i.e. from the current batch) through a round-robin mechanism. This way, the assignment problem is performed on more samples: without the queue we assign B examples to num_prototypes clusters where B is the total batch size while with the queue we assign (B + queue_length) examples to num_prototypes clusters. This is especially useful when working with small batches because it improves the precision of the assignment.

If you start using the queue too early or if you use a too large queue, this can considerably disturb training: this is because the queue members are too inconsistent. After introducing the queue the loss should be lower than what it was without the queue. On the following loss curve (30 first epochs of this script) we introduced the queue at epoch 15. We observe that it made the loss go more down.

SwAV training loss batch_size=256 during the first 30 epochs

If when introducing the queue, the loss goes up and does not decrease afterwards you should stop your training and change the queue parameters. We recommend (i) using a smaller queue, (ii) starting the queue later in training.

License

See the LICENSE file for more details.

See also

PyTorch Lightning Bolts: Implementation by the Lightning team.

SwAV-TF: A TensorFlow re-implementation.

Citation

If you find this repository useful in your research, please cite:

@article{caron2020unsupervised,
  title={Unsupervised Learning of Visual Features by Contrasting Cluster Assignments},
  author={Caron, Mathilde and Misra, Ishan and Mairal, Julien and Goyal, Priya and Bojanowski, Piotr and Joulin, Armand},
  booktitle={Proceedings of Advances in Neural Information Processing Systems (NeurIPS)},
  year={2020}
}
Owner
Meta Research
Meta Research
Scripts of Machine Learning Algorithms from Scratch. Implementations of machine learning models and algorithms using nothing but NumPy with a focus on accessibility. Aims to cover everything from basic to advance.

Algo-ScriptML Python implementations of some of the fundamental Machine Learning models and algorithms from scratch. The goal of this project is not t

Algo Phantoms 81 Nov 26, 2022
Training Certifiably Robust Neural Networks with Efficient Local Lipschitz Bounds (Local-Lip)

Training Certifiably Robust Neural Networks with Efficient Local Lipschitz Bounds (Local-Lip) Introduction TL;DR: We propose an efficient and trainabl

17 Dec 01, 2022
Code for the SIGIR 2022 paper "Hybrid Transformer with Multi-level Fusion for Multimodal Knowledge Graph Completion"

MKGFormer Code for the SIGIR 2022 paper "Hybrid Transformer with Multi-level Fusion for Multimodal Knowledge Graph Completion" Model Architecture Illu

ZJUNLP 68 Dec 28, 2022
Adversarial Self-Defense for Cycle-Consistent GANs

Adversarial Self-Defense for Cycle-Consistent GANs This is the official implementation of the CycleGAN robust to self-adversarial attacks used in pape

Dina Bashkirova 10 Oct 10, 2022
Adversarial examples to the new ConvNeXt architecture

Adversarial examples to the new ConvNeXt architecture To get adversarial examples to the ConvNeXt architecture, run the Colab: https://github.com/stan

Stanislav Fort 19 Sep 18, 2022
It's A ML based Web Site build with python and Django to find the breed of the dog

ML-Based-Dog-Breed-Identifier This is a Django Based Web Site To Identify the Breed of which your DOG belogs All You Need To Do is to Follow These Ste

Sanskar Dwivedi 2 Oct 12, 2022
[CVPR 2021] Generative Hierarchical Features from Synthesizing Images

[CVPR 2021] Generative Hierarchical Features from Synthesizing Images

GenForce: May Generative Force Be with You 148 Dec 09, 2022
A fast poisson image editing implementation that can utilize multi-core CPU or GPU to handle a high-resolution image input.

Poisson Image Editing - A Parallel Implementation Jiayi Weng (jiayiwen), Zixu Chen (zixuc) Poisson Image Editing is a technique that can fuse two imag

Jiayi Weng 110 Dec 27, 2022
This is an official implementation of CvT: Introducing Convolutions to Vision Transformers.

Introduction This is an official implementation of CvT: Introducing Convolutions to Vision Transformers. We present a new architecture, named Convolut

Bin Xiao 175 Jan 08, 2023
Storage-optimizer - Identify potintial optimizations on the cloud storage accounts

Storage Optimizer Identify potintial optimizations on the cloud storage accounts

Zaher Mousa 1 Feb 13, 2022
A smart Chat bot that can help to know about corona virus and Make prediction of corona using X-ray.

TRINIT_Hum_kuchh_nahi_karenge_ML01 Document Link https://github.com/Jatin-Goyal-552/TRINIT_Hum_kuchh_nahi_karenge_ML01/blob/main/hum_kuchh_nahi_kareng

JatinGoyal 1 Feb 03, 2022
DyStyle: Dynamic Neural Network for Multi-Attribute-Conditioned Style Editing

DyStyle: Dynamic Neural Network for Multi-Attribute-Conditioned Style Editing Figure: Joint multi-attribute edits using DyStyle model. Great diversity

74 Dec 03, 2022
Instance-wise Feature Importance in Time (FIT)

Instance-wise Feature Importance in Time (FIT) FIT is a framework for explaining time series perdiction models, by assigning feature importance to eve

Sana 46 Dec 25, 2022
Training RNNs as Fast as CNNs

News SRU++, a new SRU variant, is released. [tech report] [blog] The experimental code and SRU++ implementation are available on the dev branch which

ASAPP Research 2.1k Jan 01, 2023
Image transformations designed for Scene Text Recognition (STR) data augmentation. Published at ICCV 2021 Workshop on Interactive Labeling and Data Augmentation for Vision.

Data Augmentation for Scene Text Recognition (ICCV 2021 Workshop) (Pronounced as "strog") Paper Arxiv Why it matters? Scene Text Recognition (STR) req

Rowel Atienza 152 Dec 28, 2022
Training BERT with Compute/Time (Academic) Budget

Training BERT with Compute/Time (Academic) Budget This repository contains scripts for pre-training and finetuning BERT-like models with limited time

Intel Labs 263 Jan 07, 2023
Companion code for the paper "An Infinite-Feature Extension for Bayesian ReLU Nets That Fixes Their Asymptotic Overconfidence" (NeurIPS 2021)

ReLU-GP Residual (RGPR) This repository contains code for reproducing the following NeurIPS 2021 paper: @inproceedings{kristiadi2021infinite, title=

Agustinus Kristiadi 4 Dec 26, 2021
Code for "Unsupervised Source Separation via Bayesian inference in the latent domain"

LQVAE-separation Code for "Unsupervised Source Separation via Bayesian inference in the latent domain" Paper Samples GT Compressed Separated Drums GT

Michele Mancusi 30 Oct 25, 2022
Adversarial Robustness Toolbox (ART) - Python Library for Machine Learning Security - Evasion, Poisoning, Extraction, Inference - Red and Blue Teams

Adversarial Robustness Toolbox (ART) is a Python library for Machine Learning Security. ART provides tools that enable developers and researchers to defend and evaluate Machine Learning models and ap

3.4k Jan 04, 2023
Doing fast searching of nearest neighbors in high dimensional spaces is an increasingly important problem

Benchmarking nearest neighbors Doing fast searching of nearest neighbors in high dimensional spaces is an increasingly important problem, but so far t

Erik Bernhardsson 3.2k Jan 03, 2023