Vision foundation models

DINOv2 · 2023-04Self-distillation representations

Scales self-distillation-based self-supervised ViT training to produce general-purpose visual features that transfer without finetuning; becomes the default frozen encoder that REPA, VA-VAE, and RAE all align to or generate direct…

MAE · 2021-11Self-distillation representations

Masks most of an image's patches and trains a ViT to reconstruct the missing pixels; one of the three encoder families (with DINO, CLIP/SigLIP) RAE ablates as candidate frozen representations, and the weakest of the three by RAE's…

CLIP · 2021-02Contrastive language-image pretraining

Trains an image encoder and a text encoder jointly with a contrastive objective on 400M (image, text) pairs. Predates this atlas's 2023-2026 window and is not a diffusion paper, but its text tower became the first widely used prom…

REPA spatial-structure study · 2025-12Representation-aligned latents

Runs representation alignment against 27 different encoders and finds that what predicts a generation gain is patch-level spatial structure, not the global linear-probe accuracy the field had been using as a proxy. The most direct…

DINOv3 · 2025-08Self-distillation representations

Scales DINO-style self-supervised training to 7B parameters and adds Gram anchoring to keep dense features high-quality over very long training, fixing a known DINOv2 dense-feature degradation issue; the encoder SVG generates dire…

Perception Encoder · 2025-04Contrastive language-image pretraining

Finds that the most useful embeddings in a contrastively trained vision model sit in intermediate layers rather than at the output, and builds a recipe around extracting them. Relevant to representation-space generation because it…

SigLIP 2 · 2025-02Contrastive language-image pretraining

Extends SigLIP's sigmoid contrastive language-image objective with captioning, self-distillation and masked-prediction losses in one recipe, improving dense/localization features specifically; the frozen encoder Scale-RAE uses for…

AM-RADIO · 2023-12Agglomerative multi-teacher distillation

Distils several specialist foundation models, CLIP for semantics, DINOv2 for dense structure, SAM for segmentation, into one student backbone, on the argument that no single pretraining objective is best at everything.

ViT Registers · 2023-09Self-distillation representations

Identifies high-norm artifact tokens that appear in trained vision transformers and pollute their dense feature maps, and fixes them with a few extra register tokens that give the model somewhere to put global information. Now sta…

SigLIP · 2023-03Contrastive language-image pretraining

Replaces CLIP's softmax contrastive loss with a pairwise sigmoid loss, which removes the need for a global view of the batch and makes language-image pretraining practical at much smaller batch sizes.

I-JEPA · 2023-01Self-distillation representations

Predicts representations of masked regions in latent space rather than reconstructing their pixels, on the argument that pixel reconstruction wastes capacity on detail that carries no semantic content. Tested as a REPA alignment t…

Web-SSL · 2025-04Self-distillation representations

Trains a family of pure self-supervised (no language) vision models from 0.3B to 7B parameters on the same web data CLIP uses, and finds visual SSL keeps improving with scale where CLIP-style training saturates — evidence against …

RADIOv2.5 · 2024-12Agglomerative multi-teacher distillation

Refines the multi-teacher distillation recipe, addressing resolution and mode-switching problems that the first agglomerative models suffered.

AIMv2 · 2024-11Contrastive language-image pretraining

Pretrains a vision encoder by autoregressively predicting both image patches and paired text, a third option beside contrastive alignment and masked reconstruction, and reports clean scaling with strong frozen-backbone recognition…

UNIC · 2024-08Agglomerative multi-teacher distillation

Multi-teacher distillation with a ladder of expendable projectors and teacher-dropping regularisation. Independent evidence that agglomerating several foundation models into one backbone is a general idea rather than one company's…

Theia · 2024-07Agglomerative multi-teacher distillation

Applies the same multi-teacher distillation bet in robotics, where a single backbone that inherits several specialists is worth more than any one of them.

C-RADIOv4 · 2026-01Agglomerative multi-teacher distillation

The current agglomerative frontier, and a useful signal about consensus: its teachers are now DINOv3, SigLIP 2 and SAM 3, which is the field converging on the same three encoders that the generative-latent work also reaches for.

V-JEPA 2 · 2025-06Self-distillation representations

Carries the joint-embedding predictive idea to video and shows the resulting model supports action-conditioned prediction and robot planning. Included as a pointer: video is out of this atlas's scope, but the representation family…