SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.01571 · NeurIPS · 2022

VICRegL: Self-Supervised Learning of Local Visual Features

Yann Lecun, Adrien Bardes, Jean Ponce, Convolutional Encoder

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 20 functions out of this paper's own repositories and ran 16 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
facebookresearch/VICRegL canonical 3 of 3
lightly-ai/lightly — 5 of 7
facebookresearch/vicregl — 5 of 7
Futurne/vicreg-loss — 2 of 2
copy not recorded — 1 of 1
FunctionStatusWhere it lives
GatherLayer Ran lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("ba565bd2bd01c209")
MLP Ran facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("0393d6f458d1bd7e")
VICRegLLoss Ran Futurne/vicreg-loss/vicreg_loss/vicregl.py
pointer only (licence: NONE) · get_code("813497ec5b533d79")
VICRegLoss Ran Futurne/vicreg-loss/vicreg_loss/vicregl.py
pointer only (licence: NONE) · get_code("43e2587f6349819b")
VICRegLoss Ran lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("90c615a3b586fa8c")
accuracy Ran this paper's copy was not recorded; identical code first harvested from facebookresearch/barlowtwins
pointer only · get_code("b0f936d4d6ae3b8c")
accuracy Ran facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("f59a1c321043b890")
batch_all_gather Ran facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("e413e4905a6a548b")
batched_index_select Ran facebookresearch/VICRegL/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("1345f4fc63011f70")
covariance_loss Ran lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("15d4e6fc5c5683f1")
invariance_loss Ran lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("a20e1319851797c7")
make_inputs Ran facebookresearch/VICRegL/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("62036ceee17bee5d")
neirest_neighbores Ran facebookresearch/VICRegL/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("d53b48478f98d717")
neirest_neighbores_on_l2 Ran facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("80bfa31403ede22a")
neirest_neighbores_on_location Ran facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("8e775945d18d2299")
variance_loss Ran lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("44598859a9de393b")
VICRegL Not yet run facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("bbe9eaa0bec2c145")
VICRegLLoss Not yet run lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("6fc75c85f0b33a2c")
gather Not yet run lightly-ai/lightly/lightly/loss/vicregl_loss.py
code served (permissive licence) · get_code("f60f882797a1d561")
gather_center Not yet run facebookresearch/vicregl/main_vicregl.py
pointer only (licence: NOASSERTION) · get_code("7fe4deeb043278d7")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Most recent self-supervised methods for learning image representations focus on either producing a global feature with invariance properties, or producing a set of local features. The former works best for classification tasks while the latter is best for detection and segmentation tasks. This paper explores the fundamental tradeoff between learning local and global features. A new method called VICRegL is proposed that learns good global and local features simultaneously, yielding excellent performance on detection and segmentation tasks while maintaining good performance on classification tasks. Concretely, two identical branches of a standard convolutional net architecture are fed two differently distorted versions of the same image. The VICReg criterion is applied to pairs of global feature vectors. Simultaneously, the VICReg criterion is applied to pairs of local feature vectors occurring before the last pooling layer. Two local feature vectors are attracted to each other if their l 2 -distance is below a threshold or if their relative locations are consistent with a known geometric transformation between the two input images. We demonstrate strong performance on linear classification and segmentation transfer tasks. Code and pretrained models are publicly available at: https://github.com/facebookresearch/VICRegL

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2210.01571")
get_code_for_paper("2210.01571")
have("2210.01571")

Connect an agent — have() is free.