simclr

SimCLR

A simple framework for contrastive learning of visual representations

Abstract

This paper presents SimCLR: a simple framework for contrastive learning of visual representations. We simplify recently proposed contrastive self-supervised learning algorithms without requiring specialized architectures or a memory bank. In order to understand what enables the contrastive prediction tasks to learn useful representations, we systematically study the major components of our framework. We show that (1) composition of data augmentations plays a critical role in defining effective predictive tasks, (2) introducing a learnable nonlinear transformation between the representation and the contrastive loss substantially improves the quality of the learned representations, and (3) contrastive learning benefits from larger batch sizes and more training steps compared to supervised learning. By combining these findings, we are able to considerably outperform previous methods for self-supervised and semi-supervised learning on ImageNet. A linear classifier trained on self-supervised representations learned by SimCLR achieves 76.5% top-1 accuracy, which is a 7% relative improvement over previous state-of-the-art, matching the performance of a supervised ResNet-50.

How to use it?

Predict image

from mmpretrain import inference_model

predict = inference_model('resnet50_simclr-200e-pre_8xb512-linear-coslr-90e_in1k', 'demo/bird.JPEG')
print(predict['pred_class'])
print(predict['pred_score'])

Use the model

import torch
from mmpretrain import get_model

model = get_model('simclr_resnet50_16xb256-coslr-200e_in1k', pretrained=True)
inputs = torch.rand(1, 3, 224, 224)
out = model(inputs)
print(type(out))
# To extract features.
feats = model.extract_feat(inputs)
print(type(feats))

Train/Test Command

Prepare your dataset according to the docs.

Train:

python tools/train.py configs/simclr/simclr_resnet50_16xb256-coslr-200e_in1k.py

Test:

python tools/test.py configs/simclr/benchmarks/resnet50_8xb512-linear-coslr-90e_in1k.py https://download.openmmlab.com/mmselfsup/1.x/simclr/simclr_resnet50_16xb256-coslr-200e_in1k/resnet50_linear-8xb512-coslr-90e_in1k/resnet50_linear-8xb512-coslr-90e_in1k_20220825-f12c0457.pth

Models and results

Pretrained models

Model	Params (M)	Flops (G)	Config	Download
simclr_resnet50_16xb256-coslr-200e_in1k	27.97	4.11	config	model | log
simclr_resnet50_16xb256-coslr-800e_in1k	27.97	4.11	config	model | log

Image Classification on ImageNet-1k

Model	Pretrain	Params (M)	Flops (G)	Top-1 (%)	Config	Download
resnet50_simclr-200e-pre_8xb512-linear-coslr-90e_in1k	SIMCLR 200-Epochs	25.56	4.11	66.90	config	model | log
resnet50_simclr-800e-pre_8xb512-linear-coslr-90e_in1k	SIMCLR 800-Epochs	25.56	4.11	69.20	config	model | log

Citation

@inproceedings{chen2020simple,
  title={A simple framework for contrastive learning of visual representations},
  author={Chen, Ting and Kornblith, Simon and Norouzi, Mohammad and Hinton, Geoffrey},
  booktitle={ICML},
  year={2020},
}