Unsupervised Learning of Visual Features

by Contrasting Cluster Assignments

Mathilde Caron1,2 Ishan Misra2 Julien Mairal1

Priya Goyal2 Piotr Bojanowski2 Armand Joulin2

1
Inria∗ 2
Facebook AI Research

Abstract
Unsupervised image representations have significantly reduced the gap with su-
pervised pretraining, notably with the recent achievements of contrastive learning
methods. These contrastive methods typically work online and rely on a large num-
ber of explicit pairwise feature comparisons, which is computationally challenging.
In this paper, we propose an online algorithm, SwAV, that takes advantage of con-
trastive methods without requiring to compute pairwise comparisons. 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 as in contrastive learning. Simply put,
we use a “swapped” prediction mechanism where we predict the code 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. Compared to previous
contrastive methods, our method is more memory efficient since it does not require
a large memory bank or a special momentum network. In addition, we also propose
a new data augmentation strategy, multi-crop, that uses a mix of views with
different resolutions in place of two full-resolution views, without increasing the
memory or compute requirements. We validate our findings by achieving 75.3%
top-1 accuracy on ImageNet with ResNet-50, as well as surpassing supervised
pretraining on all the considered transfer tasks.

1 Introduction
Unsupervised visual representation learning, or self-supervised learning, aims at obtaining features
without using manual annotations and is rapidly closing the performance gap with supervised pre-
training in computer vision [9, 20, 37]. Many recent state-of-the-art methods build upon the instance
discrimination task that considers each image of the dataset (or “instance”) and its transformations as
a separate class [15]. This task yields representations that are able to discriminate between different
images, while achieving some invariance to image transformations. Recent self-supervised methods
that use instance discrimination rely on a combination of two elements: (i) a contrastive loss [19] and
(ii) a set of image transformations. The contrastive loss removes the notion of instance classes by
directly comparing image features while the image transformations define the invariances encoded in
the features. Both elements are essential to the quality of the resulting networks [9, 37] and our work
improves upon both the objective function and the transformations.
* Univ. Grenoble Alpes, Inria, CNRS, Grenoble INP, LJK, 38000 Grenoble, France
Correspondence to [email protected]
Code: https://github.com/facebookresearch/swav

34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada.
The contrastive loss explicitly compares pairs of image representations to push away representations
from different images while pulling together those from transformations, or views, of the same image.
Since computing all the pairwise comparisons on a large dataset is not practical, most implementations
approximate the loss by reducing the number of comparisons to random subsets of images during
training [9, 20, 49]. An alternative to approximate the loss is to approximate the task—that is to
relax the instance discrimination problem. For example, clustering-based methods discriminate
between groups of images with similar features instead of individual images [6]. The objective in
clustering is tractable, but it does not scale well with the dataset as it requires a pass over the entire
dataset to form image “codes” (i.e., cluster assignments) that are used as targets during training. In
this work, we use a different paradigm and propose to compute the codes online while enforcing
consistency between codes obtained from views of the same image. Comparing cluster assignments
allows to contrast different image views while not relying on explicit pairwise feature comparisons.
Specifically, we propose a simple “swapped” prediction problem where we predict the code of a
view from the representation of another view. We learn features by Swapping Assignments between
multiple Views of the same image (SwAV). The features and the codes are learned online, allowing
our method to scale to potentially unlimited amounts of data. In addition, SwAV works with small
and large batch sizes and does not need a large memory bank [49] or a momentum encoder [20].
Besides our online clustering-based method, we also propose an improvement to the image trans-
formations. Most contrastive methods compare one pair of transformations per image, even though
there is evidence that comparing more views during training improves the resulting model [37]. In
this work, we propose multi-crop that uses smaller-sized images to increase the number of views
while not increasing the memory or computational requirements during training. We also observe that
mapping small parts of a scene to more global views significantly boosts the performance. Directly
working with downsized images introduces a bias in the features [45], which can be avoided by
using a mix of different sizes. Our strategy is simple, yet effective, and can be applied to many
self-supervised methods with consistent gain in performance.
We validate our contributions by evaluating our method on several standard self-supervised bench-
marks. In particular, on the ImageNet linear evaluation protocol, we reach 75.3% top-1 accuracy with
a standard ResNet-50, and 78.5% with a wider model. We also show that our multi-crop strategy
is general, and improves the performance of different self-supervised methods, namely SimCLR [9],
DeepCluster [6], and SeLa [2], between 2% and 4% top-1 accuracy on ImageNet. Overall, we make
the following contributions:

• We propose a scalable online clustering loss that improves performance by +2% on ImageNet and
works in both large and small batch settings without a large memory bank or a momentum encoder.
• We introduce the multi-crop strategy to increase the number of views of an image with no
computational or memory overhead. We observe a consistent improvement of between 2% and 4%
on ImageNet with this strategy on several self-supervised methods.
• Combining both technical contributions into a single model, we improve the performance of self-
supervised by +4.2% on ImageNet with a standard ResNet and outperforms supervised ImageNet
pretraining on multiple downstream tasks. This is the first method to do so without finetuning the
features, i.e., only with a linear classifier on top of frozen features.

2 Related Work

Instance and contrastive learning. Instance-level classification considers each image in a dataset
as its own class [4, 15, 49]. Dosovitskiy et al. [15] assign a class explicitly to each image and learn a
linear classifier with as many classes as images in the dataset. As this approach becomes quickly
intractable, Wu et al. [49] mitigate this issue by replacing the classifier with a memory bank that stores
previously-computed representations. They rely on noise contrastive estimation [18] to compare
instances, which is a special form of contrastive learning [24, 40]. He et al. [20] improve the training
of contrastive methods by storing representations from a momentum encoder instead of the trained
network. More recently, Chen et al. [9] show that the memory bank can be entirely replaced with the
elements from the same batch if the batch is large enough. In contrast to this line of works, we avoid
comparing every pair of images by mapping the image features to a set of trainable prototype vectors.

2
Clustering for deep representation learning. Our work is also related to clustering-based meth-
ods [2, 3, 6, 7, 17, 25, 50, 53, 54, 59]. Caron et al. [6] show that k-means assignments can be used as
pseudo-labels to learn visual representations. This method scales to large uncurated dataset and can
be used for pre-training of supervised networks [7]. However, their formulation is not principled and
recently, Asano et al. [2] show how to cast the pseudo-label assignment problem as an instance of the
optimal transport problem. We consider a similar formulation to map representations to prototype
vectors, but unlike [2] we keep the soft assignment produced by the Sinkhorn-Knopp algorithm [12]
instead of approximating it into a hard assignment. Besides, unlike Caron et al. [6, 7] and Asano et
al. [2], we obtain online assignments which allows our method to scale gracefully to any dataset size.

Handcrafted pretext tasks. Many self-supervised methods manipulate the input data to extract a
supervised signal in the form of a pretext task [1, 13, 26, 28, 30, 36, 38, 41, 42, 47, 48, 57]. We refer
the reader to Jing et al. [27] for an exhaustive and detailed review of this literature. Of particular
interest, Misra and van der Maaten [37] propose to encode the jigsaw puzzle task [39] as an invariant
for contrastive learning. Jigsaw tiles are non-overlapping crops with small resolution that cover
only part (∼20%) of the entire image area. In contrast, our multi-crop strategy consists in simply
sampling multiple random crops with two different sizes: a standard size and a smaller one.

3 Method

Our goal is to learn visual features in an online fashion without supervision. To that effect, we
propose an online clustering-based self-supervised method. Typical clustering-based methods [2, 6]
are offline in the sense that they alternate between a cluster assignment step where image features of
the entire dataset are clustered, and a training step where the cluster assignments, i.e., “codes” are
predicted for different image views. Unfortunately, these methods are not suitable for online learning
as they require multiple passes over the dataset to compute the image features necessary for clustering.
In this section, we describe an alternative where we enforce consistency between codes from different
augmentations of the same image. This solution is inspired by contrastive instance learning [49] as
we do not consider the codes as a target, but only enforce consistent mapping between views of the
same image. Our method can be interpreted as a way of contrasting between multiple image views by
comparing their cluster assignments instead of their features.
More precisely, we compute a code from an augmented version of the image and predict this code
from other augmented versions of the same image. Given two image features zt and zs from two
different augmentations of the same image, we compute their codes qt and qs by matching these
features to a set of K prototypes {c1 , . . . , cK }. We then setup a “swapped” prediction problem with
the following loss function:

L(zt , zs ) = `(zt , qs ) + `(zs , qt ), (1)

where the function `(z, q) measures the fit between features z and a code q, as detailed later.
Intuitively, our method compares the features zt and zs using the intermediate codes qt and qs . If
these two features capture the same information, it should be possible to predict the code from the
other feature. A similar comparison appears in contrastive learning where features are compared
directly [49]. In Fig. 1, we illustrate the relation between contrastive learning and our method.

3.1 Online clustering

Each image xn is transformed into an augmented view xnt by applying a transformation t sampled
from the set T of image transformations. The augmented view is mapped to a vector representation by
applying a non-linear mapping fθ to xnt . The feature is then projected to the unit sphere, i.e., znt =
fθ (xnt )/kfθ (xnt )k2 . We then compute a code qnt from this feature by mapping znt to a set of
K trainable prototypes vectors, {c1 , . . . , cK }. We denote by C the matrix whose columns are the
c1 , . . . , ck . We now describe how to compute these codes and update the prototypes online.

Swapped prediction problem. The loss function in Eq. (1) has two terms that setup the “swapped”
prediction problem of predicting the code qt from the feature zs , and qs from zt . Each term represents
the cross entropy loss between the code and the probability obtained by taking a softmax of the dot

3
 Features Codes
t~T X1 fθ Z1
t~
T X1 fθ Z1 Q1

T
t~T t~
Comparison Swapped
X t~T X t~ Prototypes C
T Prediction

t~T t~
T
X2 fθ Z2 X2 fθ Z2 Q2
Features Codes

Contrastive instance learning Swapping Assignments between Views (Ours)

Figure 1: Contrastive instance learning (left) vs. SwAV (right). In contrastive learning methods
applied to instance classification, the features from different transformations of the same images are
compared directly to each other. In SwAV, we first obtain “codes” by assigning features to prototype
vectors. We then solve a “swapped” prediction problem wherein the codes obtained from one data
augmented view are predicted using the other view. Thus, SwAV does not directly compare image
features. Prototype vectors are learned along with the ConvNet parameters by backpropragation.

products of zi and all prototypes in C, i.e.,

exp τ1 z>


(k) (k) t ck
X
`(zt , qs ) = − q(k)
s log pt , where pt =P 1 > 0 .

(2)
k k0 exp τ zt ck

where τ is a temperature parameter [49]. Taking this loss over all the images and pairs of data
augmentations leads to the following loss function for the swapped prediction problem:
N
" K  >  K  > #
1 X X 1 > 1 > X znt ck X zns ck
− z Cqns + zns Cqnt − log exp − log exp .
N n=1 τ nt τ τ τ
s,t∼T k=1 k=1

This loss function is jointly minimized with respect to the prototypes C and the parameters θ of the
image encoder fθ used to produce the features (znt )n,t .

Computing codes online. In order to make our method online, we compute the codes using only
the image features within a batch. Intuitively, as the prototypes C are used across different batches,
SwAV clusters multiple instances to the prototypes. We compute codes using the prototypes C
such that all the examples in a batch are equally partitioned by the prototypes. This equipartition
constraint ensures that the codes for different images in a batch are distinct, thus preventing the
trivial solution where every image has the same code. Given B feature vectors Z = [z1 , . . . , zB ],
we are interested in mapping them to the prototypes C = [c1 , . . . , cK ]. We denote this mapping or
codes by Q = [q1 , . . . , qB ], and optimize Q to maximize the similarity between the features and the
prototypes , i.e.,
max Tr Q> C> Z + εH(Q),


(3)
Q∈Q
P
where H is the entropy function, H(Q) = − ij Qij log Qij and ε is a parameter that controls the
smoothness of the mapping. We observe that a strong entropy regularization (i.e. using a high ε)
generally leads to a trivial solution where all samples collapse into an unique representation and are
all assigned uniformely to all prototypes. Hence, in practice we keep ε low. Asano et al. [2] enforce
an equal partition by constraining the matrix Q to belong to the transportation polytope. They work
on the full dataset, and we propose to adapt their solution to work on minibatches by restricting the
transportation polytope to the minibatch:
 
1 > 1
Q = Q ∈ RK×B + | Q1B = 1K , Q 1 K = 1 B , (4)
K B
where 1K denotes the vector of ones in dimension K. These constraints enforce that on average each
B
prototype is selected at least K times in the batch.
Once a continuous solution Q∗ to Prob. (3) is found, a discrete code can be obtained by using a
rounding procedure [2]. Empirically, we found that discrete codes work well when computing codes
in an offline manner on the full dataset as in Asano et al. [2]. However, in the online setting where

4
Method Arch. Param. Top1 Supervised SwAV
Supervised R50 24 76.5
80
Colorization [56] R50 24 39.6
Jigsaw [39] R50 24 45.7 78 77.9 78.5

ImageNet top-1 accuracy

NPID [49] R50 24 54.0
BigBiGAN [14] R50 24 56.6
76 77.3 SimCLR-x4
LA [59] R50 24 58.8
74 75.3 SimCLR-x2
NPID++ [37] R50 24 59.0
MoCo [20] R50 24 60.6
SeLa [2] R50 24 61.5
PIRL [37] R50 24 63.6 72 MoCov2 CPCv2
CPC v2 [23] R50 24 63.8
PCL [31] R50 24 65.9 70 SimCLR-x1
CMC
SimCLR [9] R50 24 70.0
AMDIM
MoCov2 [10] R50 24 71.1 68 24M94M 375M 586M
SwAV R50 24 75.3 number of parameters
Figure 2: Linear classification on ImageNet. Top-1 accuracy for linear models trained on frozen
features from different self-supervised methods. (left) Performance with a standard ResNet-50.
(right) Performance as we multiply the width of a ResNet-50 by a factor ×2, ×4, and ×5.

we use only minibatches, using the discrete codes performs worse than using the continuous codes.
An explanation is that the rounding needed to obtain discrete codes is a more aggressive optimization
step than gradient updates. While it makes the model converge rapidly, it leads to a worse solution.
We thus preserve the soft code Q∗ instead of rounding it. These soft codes Q∗ are the solution of
Prob. (3) over the set Q and takes the form of a normalized exponential matrix [12]:
 > 
∗ C Z
Q = Diag(u) exp Diag(v), (5)
ε
where u and v are renormalization vectors in RK and RB respectively. The renormalization vectors
are computed using a small number of matrix multiplications using the iterative Sinkhorn-Knopp
algorithm [12]. In practice, we observe that using only 3 iterations is fast and sufficient to obtain
good performance. Indeed, this algorithm can be efficiently implemented on GPU, and the alignment
of 4K features to 3K codes takes 35ms in our experiments, see § 4.

Working with small batches. When the number B of batch features is too small compared to
the number of prototypes K, it is impossible to equally partition the batch into the K prototypes.
Therefore, when working with small batches, we use features from the previous batches to augment
the size of Z in Prob. (3). Then, we only use the codes of the batch features in our training loss. In
practice, we store around 3K features, i.e., in the same range as the number of code vectors. This
means that we only keep features from the last 15 batches with a batch size of 256, while contrastive
methods typically need to store the last 65K instances obtained from the last 250 batches [20].

3.2 Multi-crop: Augmenting views with smaller images

As noted in prior works [9, 37], comparing random crops of an image plays a central role by capturing
information in terms of relations between parts of a scene or an object. Unfortunately, increasing
the number of crops or “views” quadratically increases the memory and compute requirements. We
propose a multi-crop strategy where we use two standard resolution crops and sample V additional
low resolution crops that cover only small parts of the image. Using low resolution images ensures
only a small increase in the compute cost. Specifically, we generalize the loss of Eq (1):
X VX
+2
L(zt1 , zt2 , . . . , ztV +2 ) = 1v6=i `(ztv , qti ). (6)
i∈{1,2} v=1
Note that we compute codes using only the full resolution crops. Indeed, computing codes for all crops
increases the computational time and we observe in practice that it also alters the transfer performance

5
Table 1: Semi-supervised learning on ImageNet with a ResNet-50. We finetune the model with
1% and 10% labels and report top-1 and top-5 accuracies. *: uses RandAugment [11].
1% labels 10% labels
Method Top-1 Top-5 Top-1 Top-5
Supervised 25.4 48.4 56.4 80.4
Methods using UDA [51] - - 68.8* 88.5*
label-propagation FixMatch [44] - - 71.5* 89.1*
PIRL [37] 30.7 57.2 60.4 83.8
Methods using PCL [31] - 75.6 - 86.2
self-supervision only SimCLR [9] 48.3 75.5 65.6 87.8
SwAV 53.9 78.5 70.2 89.9

of the resulting network. An explanation is that using only partial information (small crops cover only
small area of images) degrades the assignment quality. Figure 3 shows that multi-crop improves
the performance of several self-supervised methods and is a promising augmentation strategy.

4 Main Results

We analyze the features learned by SwAV by transfer learning on multiple datasets. We implement
in SwAV the improvements used in SimCLR, i.e., LARS [55], cosine learning rate [34, 37] and the
MLP projection head [9]. We provide the full details and hyperparameters for pretraining and transfer
learning in the supplementary material.

4.1 Evaluating the unsupervised features on ImageNet

We evaluate the features of a ResNet-50 [22] trained with SwAV on ImageNet by two experiments:
linear classification on frozen features and semi-supervised learning by finetuning with few labels.
When using frozen features (Fig. 2 left), SwAV outperforms the state of the art by +4.2% top-1
accuracy and is only 1.2% below the performance of a fully supervised model. Note that we train
SwAV during 800 epochs with large batches (4096). We refer to Fig. 3 for results with shorter
trainings and to Table 3 for experiments with small batches. On semi-supervised learning (Table 1),
SwAV outperforms other self-supervised methods and is on par with state-of-the-art semi-supervised
models [44], despite the fact that SwAV is not specifically designed for semi-supervised learning.
Variants of ResNet-50. Figure 2 (right) shows the performance of multiple variants of ResNet-50
with different widths [29]. The performance of our model increases with the width of the model, and
follows a similar trend to the one obtained with supervised learning. When compared with concurrent
work like SimCLR, we see that SwAV reduces the difference with supervised models even further.
Indeed, for large architectures, our method shrinks the gap with supervised training to 0.6%.

Table 2: Transfer learning on downstream tasks. Comparison between features from ResNet-50
trained on ImageNet with SwAV or supervised learning. We consider two settings. (1) Linear
classification on top of frozen features. We report top-1 accuracy on all datasets except VOC07 where
we report mAP. (2) Object detection with finetuned features on VOC07+12 trainval using Faster
R-CNN [43] and on COCO [32] using DETR [5]. We report the most standard detection metrics for
these datasets: AP50 on VOC07+12 and AP on COCO.
Linear Classification Object Detection
Places205 VOC07 iNat18 VOC07+12 (Faster R-CNN) COCO (DETR)
Supervised 53.2 87.5 46.7 81.3 40.8
SwAV 56.7 88.9 48.6 82.6 42.1

6
Table 3: Training in small batch setting. Top-1 accuracy on ImageNet with a linear classifier
trained on top of frozen features from a ResNet-50. All methods are trained with a batch size of 256.
We also report the number of stored features, the type of cropping used and the number of epochs.
Method Mom. Encoder Stored Features multi-crop epoch batch Top-1
SimCLR 0 2×224 200 256 61.9
MoCov2 X 65, 536 2×224 200 256 67.5
MoCov2 X 65, 536 2×224 800 256 71.1
SwAV 3, 840 2×160 + 4×96 200 256 72.0
SwAV 3, 840 2×224 + 6×96 200 256 72.7
SwAV 3, 840 2×224 + 6×96 400 256 74.3

4.2 Transferring unsupervised features to downstream tasks

We test the generalization of ResNet-50 features trained with SwAV on ImageNet (without labels) by
transferring to several downstream vision tasks. In Table 2, we compare the performance of SwAV
features with ImageNet supervised pretraining. First, we report the linear classification performance
on the Places205 [58], VOC07 [16], and iNaturalist2018 [46] datasets. Our method outperforms
supervised features on all three datasets. Note that SwAV is the first self-supervised method to
surpass ImageNet supervised features on these datasets. Second, we report network finetuning on
object detection on VOC07+12 using Faster R-CNN [43] and on COCO [32] with DETR [5]. 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 use DETR because, unlike Faster
R-CNN [21], using a pretrained backbone in this framework is crucial to obtain good results compared
to training from scratch [5]. In Table 2, we show that SwAV outperforms the supervised pretrained
model on both VOC07+12 and COCO datasets. Note that this is line with previous works that also
show that self-supervision can outperform supervised pretraining on object detection [17, 20, 37].
We report more detection evaluation metrics and results from other self-supervised methods in
the supplementary material. Overall, our SwAV ResNet-50 model surpasses supervised ImageNet
pretraining on all the considered transfer tasks and datasets. We have released this model so other
researchers might also benefit by replacing the ImageNet supervised network with our model.

4.3 Training with small batches

We train SwAV with small batches of 256 images on 4 GPUs and compare with MoCov2 and
SimCLR trained in the same setup. In Table 3, we see that SwAV maintains state-of-the-art
performance even when trained in the small batch setting. Note that SwAV only stores a queue
of 3, 840 features. In comparison, to obtain good performance, MoCov2 needs to store 65, 536
features while keeping an additional momentum encoder network. When SwAV is trained using
2×160 + 4×96 crops, SwAV has a running time 1.2× higher than SimCLR with 2×224 crops and
is around 1.4× slower than MoCov2 due to the additional back-propagation [10]. Hence, one epoch
of MoCov2 or SimCLR is faster in wall clock time than one of SwAV, but these methods need
more epochs for good downstream performance. Indeed, as shown in Table 3, SwAV learns much
faster and reaches higher performance in 4× fewer epochs: 72% after 200 epochs (102 hours) while
MoCov2 needs 800 epochs to achieve 71.1%. Increasing the resolution and the number of epochs,
SwAV reaches 74.3% with a small batch size, a small number of stored features and no momentum
encoder. Finally, note that SwAV could be combined with a momentum mechanism and a large
queue [20]; we leave these explorations to future work.

5 Ablation Study
5.1 Clustering-based self-supervised learning

Improving prior clustering-based approaches. In this section, we re-implement and improve

previously published clustering-based models in order to assess if they can compete with recent
contrastive methods such as SimCLR. In particular, we consider two clustering-based models:

7
 Top-1 ∆ 76
Method 2x224 2x160+4x96 75.3

top-1 accucary
Supervised 76.5 76.0 −0.5 74 74.6 (50h)
73.9 (25h)
Contrastive-instance approaches (12h30)
SimCLR 68.2 70.6 +2.4 72 64 V-100
72.1 16Gb GPUs
Clustering-based approaches (6h15)
SeLa-v2 67.2 71.8 +4.6
70
DeepCluster-v2 70.2 74.3 +4.1 100 400 800
SwAV 70.1 74.1 +4.0 number of epochs
Figure 3: Top-1 accuracy on ImageNet with a linear classifier trained on top of frozen features from a
ResNet-50. (left) Comparison between clustering-based and contrastive instance methods and
impact of multi-crop. Self-supervised methods are trained for 400 epochs and supervised models
for 200 epochs. (right) Performance as a function of epochs. We compare SwAV models trained
with different number of epochs and report their running time based on our implementation.

DeepCluster-v2 and SeLa-v2, which are obtained by applying various training improvements in-
troduced in other self-supervised learning papers to DeepCluster [6] and SeLa [2]. Among these
improvements are the use of stronger data augmentation [9], MLP projection head [9], cosine learning
rate schedule [37], temperature parameter [49], memory bank [49], multi-clustering [2], etc. Full
implementation details can be found in the supplementary material. Besides, we also improve Deep-
Cluster model by introducing explicit comparisons to k-means centroids, which increase stability
and performance. Indeed, a main issue in DeepCluster is that there is no correspondance between
two consecutive cluster assignments. Hence, the final classification layer learned for an assignment
becomes irrelevant for the following one and thus needs to be re-initialized from scratch at each
epoch. This considerably disrupts the convnet training. In DeepCluster-v2, instead of learning a
classification layer predicting the cluster assignments, we perform explicit comparison between
features and centroids.

Comparing clustering with contrastive instance learning. In Fig. 3 (left), we make a best effort
fair comparison between clustering-based and contrastive instance (SimCLR) methods by implemen-
tating these methods with the same data augmentation, number of epochs, batch-sizes, etc. In this
setting, we observe that SwAV and DeepCluster-v2 outperform SimCLR by 2% without multi-crop
and by 3.5% with multi-crop. This suggests the learning potential of clustering-based methods
over instance classification.

Advantage of SwAV compared to DeepCluster-v2. In Fig. 3 (left), we observe that SwAV per-
forms on par with DeepCluster-v2. In addition, we train DeepCluster-v2 in SwAV best setting (800
epochs - 8 crops) and obtain 75.2% top-1 accuracy on ImageNet (versus 75.3% for SwAV). However,
unlike SwAV, DeepCluster-v2 is not online which makes it impractical for extremely large datasets
(§ 5.4). For billion scale trainings for example, a single pass on the dataset is usually performed [20].
DeepCluster-v2 cannot be trained for only one epoch since it works by performing several passes on
the dataset to regularly update centroids and cluster assignments for each image.
As a matter of fact, DeepCluster-v2 can be interpreted as a special case of our proposed swapping
mechanism: swapping is done across epochs rather than within a batch. Given a crop of an image
DeepCluster-v2 predicts the assignment of another crop, which was obtained at the previous epoch.
SwAV swaps assignments directly at the batch level and can thus work online.

5.2 Applying the multi-crop strategy to different methods

In Fig. 3 (left), we report the impact of applying our multi-crop strategy on the performance
of a selection of other methods. Details of how we apply multi-crop to SimCLR loss can be
found in the supplementary material. We see that the multi-crop strategy consistently improves
the performance for all the considered methods by a significant margin of 2−4% top-1 accuracy.
Interestingly, multi-crop seems to benefit more clustering-based methods than contrastive methods.
We note that multi-crop does not improve the supervised model.

8
 84

top-1 accuracy
Method Frozen Finetuned Weak supervision*
Random 15.0 76.5 82 No pretraining
MoCo - 77.3* 80 SwAV
SimCLR 60.4 77.2
SwAV 66.5 77.8 78
10 20 30 40
model capacity
Figure 4: Pretraining on uncurated data. Top-1 accuracy on ImageNet for pretrained models
on an uncurated set of 1B random Instagram images. (left) We compare ResNet-50 pretrained
with either SimCLR or SwAV on two downstream tasks: linear classification on frozen features or
finetuned features. (right) Performance of finetuned models as we increase the capacity of a ResNext
following [35]. The capacity is provided in billions of Mult-Add operations.
*: pretrained on a curated set of 1B Instagram images filtered with 1.5k hashtags similar to ImageNet classes.

5.3 Impact of longer training

In Fig. 3 (right), we show the impact of the number of training epochs on performance for SwAV
with multi-crop. We train separate models for 100, 200, 400 and 800 epochs and report the top-1
accuracy on ImageNet using the linear classification evaluation. We train each ResNet-50 on 64 V100
16GB GPUs and a batch size of 4096. While SwAV benefits from longer training, it already achieves
strong performance after 100 epochs, i.e., 72.1% in 6h15.

5.4 Unsupervised pretraining on a large uncurated dataset

We evaluate SwAV on random, uncurated images that have different properties from ImageNet
which allows us to test if our online clustering scheme and multi-crop augmentation work out of the
box. In particular, we pretrain SwAV on an uncurated dataset of 1 billion random public non-EU
images from Instagram. We test if SwAV can serve as a pretraining method for supervised learning.
In Fig. 4 (left), we measure the performance of ResNet-50 models when transferring to ImageNet
with frozen or finetuned features. We report the results from He et al. [20] but note that their setting
is different. They use a curated set of Instagram images, filtered by hashtags similar to ImageNet
labels [35]. We compare SwAV with a randomly initialized network and with a network pretrained on
the same data using SimCLR. We observe that SwAV maintains a similar gain of 6% over SimCLR
as when pretrained on ImageNet (Fig. 2), showing that our improvements do not depend on the data
distribution. We also see that pretraining with SwAV on random images significantly improves over
training from scratch on ImageNet (+1.3%) [7, 20]. This result is in line with Caron et al. [7] and
He et al. [20]. In Fig. 4 (right), we explore the limits of pretraining as we increase the model capacity.
We consider the variants of the ResNeXt architecture [52] as in Mahajan et al. [35]. We compare
SwAV with supervised models trained from scratch on ImageNet. For all models, SwAV outperforms
training from scratch by a significant margin showing that it can take advantage of the increased
model capacity. For reference, we also include the results from Mahajan et al. [35] obtained with a
weakly-supervised model pretrained by predicting hashtags filtered to be similar to ImageNet classes.
Interestingly, SwAV performance is strong when compared to this topline despite not using any form
of supervision or filtering of the data.

6 Discussion

Self-supervised learning is rapidly progressing compared to supervised learning, even surpassing

it on transfer learning, even though the current experimental settings are designed for supervised
learning. In particular, architectures have been designed for supervised tasks, and it is not clear if
the same models would emerge from exploring architectures with no supervision. Several recent
works have shown that exploring architectures with search [33] or pruning [8] is possible without
supervision, and we plan to evaluate the ability of our method to guide model explorations.

9
Broader Impact
This work presents a self-supervised method for learning visual representations. Self-supervised
or unsupervised learning allows training models with no annotations, nor metadata. Thus, this
work increases the field of possible applications of image features to domains where annotations
are hard to collect. For example, removing the need for annotations benefits applications where
annotations require expert knowledge, like medical imaging, or are time consuming, like fine-grained
classification. This work improves unsupervised feature learning and thus many potential downstream
applications that use visual features can benefit from it. We are uncertain of all the possible new
applications, but each application has its own merits and societal implications depending on the
intentions of the individuals using the technology.
Evaluating visual representations, whether they are supervised or self-supervised, is an open research
question. Typically used benchmarks can suffer from dataset or concept bias, and thus may reinforce
or guide future research in that direction. To mitigate this, we evaluate our work on multiple different
benchmarks and hope that future researchers also take steps in this direction.

Acknowledgement. We thank Nicolas Carion, Kaiming He, Herve Jegou, Benjamin Lefaudeux,
Thomas Lucas, Francisco Massa, Sergey Zagoruyko, and the rest of Thoth and FAIR teams for their
help and fruitful discussions. Julien Mairal was funded by the ERC grant number 714381 (SOLARIS
project) and by ANR 3IA MIAI@Grenoble Alpes (ANR-19-P3IA-0003).

References
[1] Agrawal, P., Carreira, J., Malik, J.: Learning to see by moving. In: Proceedings of the Interna-
tional Conference on Computer Vision (ICCV) (2015) 3
[2] Asano, Y.M., Rupprecht, C., Vedaldi, A.: Self-labelling via simultaneous clustering and repre-
sentation learning. International Conference on Learning Representations (ICLR) (2020) 2, 3,
4, 5, 8
[3] Bautista, M.A., Sanakoyeu, A., Tikhoncheva, E., Ommer, B.: Cliquecnn: Deep unsupervised
exemplar learning. In: Advances in Neural Information Processing Systems (NeurIPS) (2016) 3
[4] Bojanowski, P., Joulin, A.: Unsupervised learning by predicting noise. In: Proceedings of the
International Conference on Machine Learning (ICML) (2017) 2
[5] Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., Zagoruyko, S.: End-to-end object
detection with transformers. arXiv preprint arXiv:2005.12872 (2020) 6, 7
[6] Caron, M., Bojanowski, P., Joulin, A., Douze, M.: Deep clustering for unsupervised learning
of visual features. In: Proceedings of the European Conference on Computer Vision (ECCV)
(2018) 2, 3, 8
[7] Caron, M., Bojanowski, P., Mairal, J., Joulin, A.: Unsupervised pre-training of image features
on non-curated data. In: Proceedings of the International Conference on Computer Vision
(ICCV) (2019) 3, 9
[8] Caron, M., Morcos, A., Bojanowski, P., Mairal, J., Joulin, A.: Pruning convolutional neural
networks with self-supervision. arXiv preprint arXiv:2001.03554 (2020) 9
[9] Chen, T., Kornblith, S., Norouzi, M., Hinton, G.: A simple framework for contrastive learning
of visual representations. arXiv preprint arXiv:2002.05709 (2020) 1, 2, 5, 6, 8
[10] Chen, X., Fan, H., Girshick, R., He, K.: Improved baselines with momentum contrastive
learning. arXiv preprint arXiv:2003.04297 (2020) 5, 7
[11] Cubuk, E.D., Zoph, B., Shlens, J., Le, Q.V.: Randaugment: Practical data augmentation with no
separate search. arXiv preprint arXiv:1909.13719 (2019) 6
[12] Cuturi, M.: Sinkhorn distances: Lightspeed computation of optimal transport. In: Advances in
Neural Information Processing Systems (NeurIPS) (2013) 3, 5
[13] Doersch, C., Gupta, A., Efros, A.A.: Unsupervised visual representation learning by context
prediction. In: Proceedings of the International Conference on Computer Vision (ICCV) (2015)
3

10
[14] Donahue, J., Simonyan, K.: Large scale adversarial representation learning. In: Advances in
Neural Information Processing Systems (NeurIPS) (2019) 5
[15] Dosovitskiy, A., Fischer, P., Springenberg, J.T., Riedmiller, M., Brox, T.: Discriminative
unsupervised feature learning with exemplar convolutional neural networks. IEEE transactions
on pattern analysis and machine intelligence 38(9), 1734–1747 (2016) 1, 2
[16] Everingham, M., Van Gool, L., Williams, C.K., Winn, J., Zisserman, A.: The pascal visual
object classes (voc) challenge. International journal of computer vision 88(2), 303–338 (2010)
7
[17] Gidaris, S., Bursuc, A., Komodakis, N., Pérez, P., Cord, M.: Learning representations by
predicting bags of visual words. arXiv preprint arXiv:2002.12247 (2020) 3, 7
[18] Gutmann, M., Hyvärinen, A.: Noise-contrastive estimation: A new estimation principle for
unnormalized statistical models. In: Proceedings of the Thirteenth International Conference on
Artificial Intelligence and Statistics (2010) 2
[19] Hadsell, R., Chopra, S., LeCun, Y.: Dimensionality reduction by learning an invariant mapping.
In: Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR) (2006)
1
[20] He, K., Fan, H., Wu, Y., Xie, S., Girshick, R.: Momentum contrast for unsupervised visual
representation learning. arXiv preprint arXiv:1911.05722 (2019) 1, 2, 5, 7, 8, 9
[21] He, K., Girshick, R., Dollár, P.: Rethinking imagenet pre-training. In: Proceedings of the
International Conference on Computer Vision (ICCV) (2019) 7
[22] He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Pro-
ceedings of the Conference on Computer Vision and Pattern Recognition (CVPR) (2016)
6
[23] Hénaff, O.J., Razavi, A., Doersch, C., Eslami, S., Oord, A.v.d.: Data-efficient image recognition
with contrastive predictive coding. arXiv preprint arXiv:1905.09272 (2019) 5
[24] Hjelm, R.D., Fedorov, A., Lavoie-Marchildon, S., Grewal, K., Bachman, P., Trischler, A.,
Bengio, Y.: Learning deep representations by mutual information estimation and maximization.
International Conference on Learning Representations (ICLR) (2019) 2
[25] Huang, J., Dong, Q., Gong, S.: Unsupervised deep learning by neighbourhood discovery. In:
Proceedings of the International Conference on Machine Learning (ICML) (2019) 3
[26] Jenni, S., Favaro, P.: Self-supervised feature learning by learning to spot artifacts. In: Pro-
ceedings of the Conference on Computer Vision and Pattern Recognition (CVPR) (2018)
3
[27] Jing, L., Tian, Y.: Self-supervised visual feature learning with deep neural networks: A survey.
arXiv preprint arXiv:1902.06162 (2019) 3
[28] Kim, D., Cho, D., Yoo, D., Kweon, I.S.: Learning image representations by completing damaged
jigsaw puzzles. In: Winter Conference on Applications of Computer Vision (WACV) (2018) 3
[29] Kolesnikov, A., Zhai, X., Beyer, L.: Revisiting self-supervised visual representation learning. In:
Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR) (2019) 6
[30] Larsson, G., Maire, M., Shakhnarovich, G.: Learning representations for automatic colorization.
In: Proceedings of the European Conference on Computer Vision (ECCV) (2016) 3
[31] Li, J., Zhou, P., Xiong, C., Socher, R., Hoi, S.C.: Prototypical contrastive learning of unsuper-
vised representations. arXiv preprint arXiv:2005.04966 (2020) 5, 6
[32] Lin, T.Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., Zitnick, C.L.:
Microsoft coco: Common objects in context. In: Proceedings of the European Conference on
Computer Vision (ECCV) (2014) 6, 7
[33] Liu, C., Dollár, P., He, K., Girshick, R., Yuille, A., Xie, S.: Are labels necessary for neural
architecture search? arXiv preprint arXiv:2003.12056 (2020) 9
[34] Loshchilov, I., Hutter, F.: Sgdr: Stochastic gradient descent with warm restarts. arXiv preprint
arXiv:1608.03983 (2016) 6

11
[35] Mahajan, D., Girshick, R., Ramanathan, V., He, K., Paluri, M., Li, Y., Bharambe, A., van der
Maaten, L.: Exploring the limits of weakly supervised pretraining. In: Proceedings of the
European Conference on Computer Vision (ECCV) (2018) 9
[36] Mahendran, A., Thewlis, J., Vedaldi, A.: Cross pixel optical flow similarity for self-supervised
learning. arXiv preprint arXiv:1807.05636 (2018) 3
[37] Misra, I., van der Maaten, L.: Self-supervised learning of pretext-invariant representations.
arXiv preprint arXiv:1912.01991 (2019) 1, 2, 3, 5, 6, 7, 8
[38] Misra, I., Zitnick, C.L., Hebert, M.: Shuffle and learn: unsupervised learning using temporal
order verification. In: Proceedings of the European Conference on Computer Vision (ECCV)
(2016) 3
[39] Noroozi, M., Favaro, P.: Unsupervised learning of visual representations by solving jigsaw
puzzles. In: Proceedings of the European Conference on Computer Vision (ECCV) (2016) 3, 5
[40] Oord, A.v.d., Li, Y., Vinyals, O.: Representation learning with contrastive predictive coding.
arXiv preprint arXiv:1807.03748 (2018) 2
[41] Pathak, D., Girshick, R., Dollár, P., Darrell, T., Hariharan, B.: Learning features by watching
objects move. In: Proceedings of the Conference on Computer Vision and Pattern Recognition
(CVPR) (2017) 3
[42] Pathak, D., Krahenbuhl, P., Donahue, J., Darrell, T., Efros, A.A.: Context encoders: Feature
learning by inpainting. In: Proceedings of the Conference on Computer Vision and Pattern
Recognition (CVPR) (2016) 3
[43] Ren, S., He, K., Girshick, R., Sun, J.: Faster r-cnn: Towards real-time object detection with
region proposal networks. In: Advances in Neural Information Processing Systems (NeurIPS)
(2015) 6, 7
[44] Sohn, K., Berthelot, D., Li, C.L., Zhang, Z., Carlini, N., Cubuk, E.D., Kurakin, A., Zhang, H.,
Raffel, C.: Fixmatch: Simplifying semi-supervised learning with consistency and confidence.
arXiv preprint arXiv:2001.07685 (2020) 6
[45] Touvron, H., Vedaldi, A., Douze, M., Jégou, H.: Fixing the train-test resolution discrepancy. In:
Advances in Neural Information Processing Systems (NeurIPS) (2019) 2
[46] Van Horn, G., Mac Aodha, O., Song, Y., Cui, Y., Sun, C., Shepard, A., Adam, H., Perona, P.,
Belongie, S.: The inaturalist species classification and detection dataset. In: Proceedings of the
Conference on Computer Vision and Pattern Recognition (CVPR) (2018) 7
[47] Wang, X., Gupta, A.: Unsupervised learning of visual representations using videos. In: Pro-
ceedings of the International Conference on Computer Vision (ICCV) (2015) 3
[48] Wang, X., He, K., Gupta, A.: Transitive invariance for self-supervised visual representation
learning. In: Proceedings of the International Conference on Computer Vision (ICCV) (2017) 3
[49] Wu, Z., Xiong, Y., Yu, S.X., Lin, D.: Unsupervised feature learning via non-parametric instance
discrimination. In: Proceedings of the Conference on Computer Vision and Pattern Recognition
(CVPR) (2018) 2, 3, 4, 5, 8
[50] Xie, J., Girshick, R., Farhadi, A.: Unsupervised deep embedding for clustering analysis. In:
Proceedings of the International Conference on Machine Learning (ICML) (2016) 3
[51] Xie, Q., Dai, Z.D., Hovy, E., Luong, M.T., Le, Q.V.: Unsupervised data augmentation for
consistency training. arXiv preprint arXiv:1904.12848 (2020) 6
[52] Xie, S., Girshick, R., Dollár, P., Tu, Z., He, K.: Aggregated residual transformations for
deep neural networks. In: Proceedings of the Conference on Computer Vision and Pattern
Recognition (CVPR) (2017) 9
[53] Yan, X., Misra, Ishan, I., Gupta, A., Ghadiyaram, D., Mahajan, D.: ClusterFit: Improving
generalization of visual representations. In: Proceedings of the Conference on Computer Vision
and Pattern Recognition (CVPR) (2020) 3
[54] Yang, J., Parikh, D., Batra, D.: Joint unsupervised learning of deep representations and image
clusters. In: Proceedings of the Conference on Computer Vision and Pattern Recognition
(CVPR) (2016) 3
[55] You, Y., Gitman, I., Ginsburg, B.: Large batch training of convolutional networks. arXiv preprint
arXiv:1708.03888 (2017) 6

12
[56] Zhang, R., Isola, P., Efros, A.A.: Colorful image colorization. In: Proceedings of the European
Conference on Computer Vision (ECCV) (2016) 5
[57] Zhang, R., Isola, P., Efros, A.A.: Split-brain autoencoders: Unsupervised learning by cross-
channel prediction. In: Proceedings of the Conference on Computer Vision and Pattern Recog-
nition (CVPR) (2017) 3
[58] Zhou, B., Lapedriza, A., Xiao, J., Torralba, A., Oliva, A.: Learning deep features for scene
recognition using places database. In: Advances in Neural Information Processing Systems
(NeurIPS) (2014) 7
[59] Zhuang, C., Zhai, A.L., Yamins, D.: Local aggregation for unsupervised learning of visual
embeddings. In: Proceedings of the International Conference on Computer Vision (ICCV)
(2019) 3, 5

13