SkillAgentSearch skills...

EarthMiss

EarthMiss: A multimodal (Optical/SAR) land-cover dataset across 13 global cities + MetaRS code for missing-modality mapping.

Install / Use

/learn @Yi-Heng/EarthMiss
About this skill

Quality Score

0/100

Supported Platforms

Universal

README

EarthMiss

EarthMiss: A multimodal (Optical/SAR) land-cover dataset across 13 global cities + MetaRS code for missing-modality mapping.

Project homepage: https://rsidea.whu.edu.cn/EarthMiss.html

Installation

pip install --upgrade git+https://github.com/Z-Zheng/SimpleCV.git
pip install --upgrade git+https://github.com/Z-Zheng/ever.git

Dataset

The EarthMiss dataset is released at Zenodo and Baidu Drive Code: jfsq

Test and Train

The data split is in ./configs/metadata/EarthMiss.py

  1. Download the dataset and update the root_path in ./config/metadata/EarthMiss.py to your local dataset directory.

  2. Download the pretrained checkpoint:
    📥 Google Drive

  3. Run evaluation:

sh ./scripts/eval.sh
sh ./scripts/train.sh

Related Skills

View on GitHub
GitHub Stars17
CategoryDevelopment
Updated24d ago
Forks1

Languages

Python

Security Score

75/100

Audited on Mar 11, 2026

No findings