AIbase
EN
Home
Vit Base Patch16 224 In21k Finetuned Eurosat
V

Vit Base Patch16 224 In21k Finetuned Eurosat

Developed by ezzouhri
Image classification model based on Google Vision Transformer (ViT) architecture, fine-tuned on the EuroSAT dataset
Image Classification
Transformers
Open Source License:Apache-2.0#Remote Sensing Image Classification#High-Precision ViT Model#EuroSAT Dataset
Downloads 30
Release Time : 10/17/2022

Model Overview

This model is a fine-tuned version of Google's pre-trained ViT model on the EuroSAT remote sensing image dataset, suitable for satellite image classification tasks

Model Features

High-Performance Image Classification
Achieves 90.22% accuracy on the EuroSAT evaluation set
Transformer-Based Architecture
Uses Vision Transformer architecture, independent of traditional CNN structures
Efficient Inference
Evaluation speed reaches 21.486 samples per second

Model Capabilities

Satellite Image Classification
Remote Sensing Image Analysis
Multi-Class Image Recognition

Use Cases

Remote Sensing Applications
Land Use Classification
Identify different land types in satellite images
Achieves 90.22% accuracy on the EuroSAT dataset
Environmental Monitoring
Analyze environmental factors such as vegetation coverage and water body changes
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase