lightly 1.4.22

A deep learning package for self-supervised learning

Lightly SSL is a computer vision framework for self-supervised learning.

• Documentation
• Github
• Discord (We have weekly paper sessions!)

We've also built a whole platform on top, with additional features for active learning and data curation. If you're interested in the Lightly Worker Solution to easily process millions of samples and run powerful algorithms on your data, check out lightly.ai. It's free to get started!

Features

This self-supervised learning framework offers the following features:

• Modular framework, which exposes low-level building blocks such as loss functions and model heads.
• Easy to use and written in a PyTorch like style.
• Supports custom backbone models for self-supervised pre-training.
• Support for distributed training using PyTorch Lightning.

Supported Models

You can find sample code for all the supported models here. We provide PyTorch, PyTorch Lightning, and PyTorch Lightning distributed examples for all models to kickstart your project.

Models:

• Barlow Twins, 2021 paper docs
• BYOL, 2020 paper docs
• DCL & DCLW, 2021 paper docs
• DINO, 2021 paper docs
• MAE, 2021 paper docs
• MSN, 2022 paper docs
• MoCo, 2019 paper docs
• NNCLR, 2021 paper docs
• PMSN, 2022 paper docs
• SimCLR, 2020 paper docs
• SimMIM, 2021 paper docs
• SimSiam, 2021 paper docs
• SMoG, 2022 paper docs
• SwaV, 2020 paper docs
• TiCo, 2022 paper docs
• VICReg, 2022 paper docs
• VICRegL, 2022 paper docs

Tutorials

Want to jump to the tutorials and see Lightly in action?

• Train MoCo on CIFAR-10
• Train SimCLR on Clothing Data
• Train SimSiam on Satellite Images
• Use Lightly with Custom Augmentations
• Pre-train a Detectron2 Backbone with Lightly

Community and partner projects:

• On-Device Deep Learning with Lightly on an ARM microcontroller

Quick Start

Lightly requires Python 3.6+ but we recommend using Python 3.7+. We recommend installing Lightly in a Linux or OSX environment.

Dependencies

• PyTorch
• Torchvision
• PyTorch Lightning v1.5+

Lightly is compatible with PyTorch and PyTorch Lightning v2.0+!

Vision transformer based models require Torchvision v0.12+.

Installation

You can install Lightly and its dependencies from PyPI with:

pip3 install lightly

We strongly recommend that you install Lightly in a dedicated virtualenv, to avoid conflicting with your system packages.

Lightly in Action

With Lightly, you can use the latest self-supervised learning methods in a modular way using the full power of PyTorch. Experiment with different backbones, models, and loss functions. The framework has been designed to be easy to use from the ground up. Find more examples in our docs.

import torch
import torchvision

from lightly import loss
from lightly import transforms
from lightly.data import LightlyDataset
from lightly.models.modules import heads


# Create a PyTorch module for the SimCLR model.
class SimCLR(torch.nn.Module):
    def __init__(self, backbone):
        super().__init__()
        self.backbone = backbone
        self.projection_head = heads.SimCLRProjectionHead(
            input_dim=512,  # Resnet18 features have 512 dimensions.
            hidden_dim=512,
            output_dim=128,
        )

    def forward(self, x):
        features = self.backbone(x).flatten(start_dim=1)
        z = self.projection_head(features)
        return z


# Use a resnet backbone.
backbone = torchvision.models.resnet18()
# Ignore the classification head as we only want the features.
backbone.fc = torch.nn.Identity()

# Build the SimCLR model.
model = SimCLR(backbone)

# Prepare transform that creates multiple random views for every image.
transform = transforms.SimCLRTransform(input_size=32, cj_prob=0.5)


# Create a dataset from your image folder.
dataset = data.LightlyDataset(input_dir="./my/cute/cats/dataset/", transform=transform)

# Build a PyTorch dataloader.
dataloader = torch.utils.data.DataLoader(
    dataset,  # Pass the dataset to the dataloader.
    batch_size=128,  # A large batch size helps with the learning.
    shuffle=True,  # Shuffling is important!
)

# Lightly exposes building blocks such as loss functions.
criterion = loss.NTXentLoss(temperature=0.5)

# Get a PyTorch optimizer.
optimizer = torch.optim.SGD(model.parameters(), lr=0.1, weight_decay=1e-6)

# Train the model.
for epoch in range(10):
    for (view0, view1), targets, filenames in dataloader:
        z0 = model(view0)
        z1 = model(view1)
        loss = criterion(z0, z1)
        loss.backward()
        optimizer.step()
        optimizer.zero_grad()
        print(f"loss: {loss.item():.5f}")

You can easily use another model like SimSiam by swapping the model and the loss function.

# PyTorch module for the SimSiam model.
class SimSiam(torch.nn.Module):
    def __init__(self, backbone):
        super().__init__()
        self.backbone = backbone
        self.projection_head = heads.SimSiamProjectionHead(512, 512, 128)
        self.prediction_head = heads.SimSiamPredictionHead(128, 64, 128)

    def forward(self, x):
        features = self.backbone(x).flatten(start_dim=1)
        z = self.projection_head(features)
        p = self.prediction_head(z)
        z = z.detach()
        return z, p


model = SimSiam(backbone)

# Use the SimSiam loss function.
criterion = loss.NegativeCosineSimilarity()

You can find a more complete example for SimSiam here.

Use PyTorch Lightning to train the model:

from pytorch_lightning import LightningModule, Trainer

class SimCLR(LightningModule):
    def __init__(self):
        super().__init__()
        resnet = torchvision.models.resnet18()
        resnet.fc = torch.nn.Identity()
        self.backbone = resnet
        self.projection_head = heads.SimCLRProjectionHead(512, 512, 128)
        self.criterion = loss.NTXentLoss()

    def forward(self, x):
        features = self.backbone(x).flatten(start_dim=1)
        z = self.projection_head(features)
        return z

    def training_step(self, batch, batch_index):
        (view0, view1), _, _ = batch
        z0 = self.forward(view0)
        z1 = self.forward(view1)
        loss = self.criterion(z0, z1)
        return loss

    def configure_optimizers(self):
        optim = torch.optim.SGD(self.parameters(), lr=0.06)
        return optim


model = SimCLR()
trainer = Trainer(max_epochs=10, devices=1, accelerator="gpu")
trainer.fit(model, dataloader)

See our docs for a full PyTorch Lightning example.

Or train the model on 4 GPUs:

# Use distributed version of loss functions.
criterion = loss.NTXentLoss(gather_distributed=True)

trainer = Trainer(
    max_epochs=10,
    devices=4,
    accelerator="gpu",
    strategy="ddp",
    sync_batchnorm=True,
    use_distributed_sampler=True,  # or replace_sampler_ddp=True for PyTorch Lightning <2.0
)
trainer.fit(model, dataloader)

We provide multi-GPU training examples with distributed gather and synchronized BatchNorm. Have a look at our docs regarding distributed training.

Benchmarks

Implemented models and their performance on various datasets. Hyperparameters are not tuned for maximum accuracy. For detailed results and more info about the benchmarks click here.

Imagenet

The following experiments have been conducted on a system with 2x4090 GPUs. Training a model takes around 4 days for 100 epochs (35 min per epoch), including kNN, linear probing, and fine-tuning evaluation.

Note: Evaluation settings are based on these papers:

• Linear: SimCLR
• Finetune: SimCLR
• KNN: InstDisc

See the benchmarking scripts for details.

Model	Backbone	Batch Size	Epochs	Linear Top1	Finetune Top1	KNN Top1	Tensorboard	Checkpoint
BarlowTwins	Res50	256	100	62.9	72.6	45.6	link	link
BYOL	Res50	256	100	62.4	74.0	45.6	link	link
DINO	Res50	128	100	68.2	72.5	49.9	link	link
SimCLR*	Res50	256	100	63.2	73.9	44.8	link	link
SimCLR* + DCL	Res50	256	100	65.1	73.5	49.6	link	link
SimCLR* + DCLW	Res50	256	100	64.5	73.2	48.5	link	link
SwAV	Res50	256	100	67.2	75.4	49.5	link	link
VICReg	Res50	256	100	63.0	73.7	46.3	link	link

*We use square root learning rate scaling instead of linear scaling as it yields better results for smaller batch sizes. See Appendix B.1 in SimCLR paper.

ImageNette

Model	Backbone	Batch Size	Epochs	KNN Top1
BarlowTwins	Res18	256	800	0.852
BYOL	Res18	256	800	0.887
DCL	Res18	256	800	0.861
DCLW	Res18	256	800	0.865
DINO	Res18	256	800	0.888
FastSiam	Res18	256	800	0.873
MAE	ViT-S	256	800	0.610
MSN	ViT-S	256	800	0.828
Moco	Res18	256	800	0.874
NNCLR	Res18	256	800	0.884
PMSN	ViT-S	256	800	0.822
SimCLR	Res18	256	800	0.889
SimMIM	ViT-B32	256	800	0.343
SimSiam	Res18	256	800	0.872
SwaV	Res18	256	800	0.902
SwaVQueue	Res18	256	800	0.890
SMoG	Res18	256	800	0.788
TiCo	Res18	256	800	0.856
VICReg	Res18	256	800	0.845
VICRegL	Res18	256	800	0.778

Cifar10

Model	Backbone	Batch Size	Epochs	KNN Top1
BarlowTwins	Res18	512	800	0.859
BYOL	Res18	512	800	0.910
DCL	Res18	512	800	0.874
DCLW	Res18	512	800	0.871
DINO	Res18	512	800	0.848
FastSiam	Res18	512	800	0.902
Moco	Res18	512	800	0.899
NNCLR	Res18	512	800	0.892
SimCLR	Res18	512	800	0.879
SimSiam	Res18	512	800	0.904
SwaV	Res18	512	800	0.884
SMoG	Res18	512	800	0.800

Terminology

Below you can see a schematic overview of the different concepts in the package. The terms in bold are explained in more detail in our documentation.

Next Steps

Head to the documentation and see the things you can achieve with Lightly!

Development

To install dev dependencies (for example to contribute to the framework) you can use the following command:

pip3 install -e ".[dev]"

For more information about how to contribute have a look here.

Running Tests

Unit tests are within the tests directory and we recommend running them using pytest. There are two test configurations available. By default, only a subset will be run:

make test-fast

To run all tests (including the slow ones) you can use the following command:

make test

To test a specific file or directory use:

pytest <path to file or directory>

Code Formatting

To format code with black and isort run:

make format

Further Reading

Self-Supervised Learning:

• Have a look at our #papers channel on discord for the newest self-supervised learning papers.
• A Cookbook of Self-Supervised Learning, 2023
• Masked Autoencoders Are Scalable Vision Learners, 2021
• Emerging Properties in Self-Supervised Vision Transformers, 2021
• Unsupervised Learning of Visual Features by Contrasting Cluster Assignments, 2021
• What Should Not Be Contrastive in Contrastive Learning, 2020
• A Simple Framework for Contrastive Learning of Visual Representations, 2020
• Momentum Contrast for Unsupervised Visual Representation Learning, 2020

FAQ

• Why should I care about self-supervised learning? Aren't pre-trained models from ImageNet much better for transfer learning?

  • Self-supervised learning has become increasingly popular among scientists over the last years because the learned representations perform extraordinarily well on downstream tasks. This means that they capture the important information in an image better than other types of pre-trained models. By training a self-supervised model on your dataset, you can make sure that the representations have all the necessary information about your images.

• How can I contribute?

  • Create an issue if you encounter bugs or have ideas for features we should implement. You can also add your own code by forking this repository and creating a PR. More details about how to contribute with code is in our contribution guide.

• Is this framework for free?

  • Yes, this framework is completely free to use and we provide the source code. We believe that we need to make training deep learning models more data efficient to achieve widespread adoption. One step to achieve this goal is by leveraging self-supervised learning. The company behind Lightly is committed to keep this framework open-source.

• If this framework is free, how is the company behind Lightly making money?

  • Training self-supervised models is only one part of our solution. The company behind Lightly focuses on processing and analyzing embeddings created by self-supervised models. By building, what we call a self-supervised active learning loop we help companies understand and work with their data more efficiently. As the Lightly Solution is a freemium product, you can try it out for free. However, we will charge for some features.
  • In any case this framework will always be free to use, even for commercial purposes.

Lightly in Research

• Reverse Engineering Self-Supervised Learning, 2023
• Learning Visual Representations via Language-Guided Sampling, 2023
• Self-Supervised Learning Methods for Label-Efficient Dental Caries Classification, 2022
• DPCL: Constrative Representation Learning with Differential Privacy, 2022
• Decoupled Contrastive Learning, 2021
• solo-learn: A Library of Self-supervised Methods for Visual Representation Learning, 2021

Company behind this Open Source Framework

Lightly is a spin-off from ETH Zurich that helps companies build efficient active learning pipelines to select the most relevant data for their models.

You can find out more about the company and it's services by following the links below:

• Homepage
• Web-App
• Lightly Solution Documentation (Lightly Worker & API)

BibTeX

If you want to cite the framework feel free to use this:

@article{susmelj2020lightly,
  title={Lightly},
  author={Igor Susmelj and Matthias Heller and Philipp Wirth and Jeremy Prescott and Malte Ebner et al.},
  journal={GitHub. Note: https://github.com/lightly-ai/lightly},
  year={2020}
}