PyTorch code and models for the DINOv2 self-supervised learning method.

All-in-one training for vision models (YOLO, ViTs, RT-DETR, DINOv3): pretraining, fine-tuning, distillation.

A microframework on top of PyTorch with first-class citizen APIs for foundation model adaptation

Vision Foundation Models for Medical AI, including RETFound, DINOv2, DINOv3

[ICLR'24 & IJCV‘25] Matcher: Segment Anything with One Shot Using All-Purpose Feature Matching

Testing adaptation of the DINOv2/3 encoders for vision tasks with Low-Rank Adaptation (LoRA)

[ECCV 2024] Improving 2D Feature Representations by 3D-Aware Fine-Tuning

[ICCV 2023 R6D] PyTorch implementation of CNOS: A Strong Baseline for CAD-based Novel Object Segmentation based on Segmenting Anything and DINOv2

Downstream-Dino-V2: A GitHub repository featuring an easy-to-use implementation of the DINOv2 model by Facebook for downstream tasks such as Classification, Semantic Segmentation and Monocular depth estimation.

Unofficial implementation of the paper "The Chosen One: Consistent Characters in Text-to-Image Diffusion Models"

框架核心是两阶段“粗筛-精滤”数据清洗流程。首先，利用CLIP的多门控决策逻辑进行宏观粗筛，精准剔除插画、图表等非摄影类噪声。随后，利用DINOv2的细粒度特征，创新采用“相对边际分数”识别处于类别边界的混淆样本，并结合GMM模型为各类别动态确定清洗标准。整个流程内置最小样本保留机制，以保护长尾分布下的稀有类别。 在清洗后的高纯度数据上，我们采用非对称协同训练策略。通过Swin Transformer与ConvNeXt双网络，利用其结构差异形成互补的错误识别模式，高效实现相互监督与纠错。同时，我们创新地对模型达成共识的样本目标化应用Balanced Softmax Loss，在精准应对长尾分布的同时，避免了放大噪声的影响。

[WACV2025] AnomalyDINO: Boosting Patch-based Few-shot Anomaly Detection with DINOv2

[CVPR'24] NeRF On-the-go: Exploiting Uncertainty for Distractor-free NeRFs in the Wild

[ICCV 2025] Official repository of the paper "Talking to DINO: Bridging Self-Supervised Vision Backbones with Language for Open-Vocabulary Segmentation"

Welcome to the project repository for POPE (Promptable Pose Estimation), a state-of-the-art technique for 6-DoF pose estimation of any object in any scene using a single reference.

Our idea is to combine the power of computer vision model and LLMs. We use YOLO, CLIP and DINOv2 to extract high-level features from images. We pass the prompt, along with the extracted features, to LLM, allowing for advanced image dataset queries.

Multi-modal AI-generated content detection: image, video, and audio. Benchmarks, training code (DINOv2, DINOv3, ReStraV, BreathNet), and evaluation pipeline for real vs. synthetic classification with calibration-aware metrics.

DINOv2 inference engine written in C/C++ using ggml and OpenCV.

Integrating SAM2 with DINOv2/v3 for segmentation

A cli program of image retrieval using dinov2

Blood Cell Foundation Model based on DINOv2

Evaluating DINOv2 for Radiology

Finetuning DINOv2 (https://github.com/facebookresearch/dinov2) on your own dataset

[NeurIPS'24] A Simple Image Segmentation Framework via In-Context Examples

PyTorch code for FLD (Feature Likelihood Divergence), FID, KID, Precision, Recall, etc. using DINOv2, InceptionV3, CLIP, etc.

[WACV 2026] Official implementation of the paper: “CountingDINO: A Training-free Pipeline for Exemplar-based Class-Agnostic Counting”

[CVPR'26] Official Code for “V²-SAM: Marrying SAM2 with Multi-Prompt Experts for Cross-View Object Correspondence”

This repository provides a 3D implementation of DINOv2 for self-supervised pretraining on volumetric (3D) medical images using Lightly, MONAI, and Pytorch Lightning!

Training and testing of DINOv2 for segmentation downstream

Visualization of the PCA as shown in Figure 1.