D

Data2vec Vision Base

Developed by facebook
Data2Vec-Vision is a self-supervised learning model based on the BEiT architecture, pretrained on the ImageNet-1k dataset, suitable for image classification tasks.
Downloads 427
Release Time : 4/14/2022

Model Overview

This model adopts the data2vec framework, learning latent representations of input data through self-distillation, supporting computer vision tasks.

Model Features

General self-supervised learning framework
Employs the unified data2vec framework applicable to multiple modalities including speech, vision, and language.
Contextual latent representation prediction
Predicts latent representations of complete input data rather than local features, capturing richer contextual information.
ImageNet pretraining
Pretrained on the ImageNet-1k dataset containing 1.2 million images, equipped with powerful visual feature extraction capabilities.

Model Capabilities

Image feature extraction
Image classification

Use Cases

Computer vision
Image classification
Classifies input images, supporting 1000 ImageNet categories.
Achieves or approaches state-of-the-art performance on multiple image classification benchmarks.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase