Florence-2-large-DOTA-v1.0-lmmrotate Open-source Model - Empowering Rotated Object Detection and Analysis in Aerial Images

Florence 2 Large DOTA V1.0 Lmmrotate

Developed by Qingyun

LMMRotate is a fine-tuned large multimodal language model specifically designed for rotating object detection tasks, particularly suitable for aerial image analysis.

Image-to-Text

TensorBoard

EnglishOpen Source License:MIT #Aerial rotating object detection #Multimodal language model #Remote sensing image analysis

Downloads 17

Release Time : 1/10/2025

Model Overview

LMMRotate is fine-tuned based on the Florence-2-large model, specifically for rotating object detection tasks in aerial images. It provides a simple baseline for aerial detection, applicable in earth science and remote sensing fields.

Model Features

Aerial image analysis

Optimized specifically for aerial images, capable of effectively handling rotating object detection tasks.

Multimodal capability

Combines visual and language understanding abilities to generate relevant text descriptions from images.

Simple baseline

Provides a straightforward yet effective baseline model for aerial detection, facilitating subsequent research and application development.

Model Capabilities

Aerial image analysis

Rotating object detection

Image-text generation

Remote sensing data processing

Use Cases

Earth science

Aerial object detection

Used to detect specific objects in aerial images, such as buildings, vehicles, etc.

Performs well on the DOTA-v1.0 dataset

Remote sensing

Geographic information extraction

Extracts geographic information features from remote sensing images.

🚀 LMMRotate 🎮: A Simple Aerial Detection Baseline of Multimodal Language Models

LMMRotate is a technical practice for fine - tuning Large Multimodal language Models for oriented object detection. It provides a simple aerial detection baseline for multimodal language models.

🚀 Quick Start

This repo stores the checkpoint of Florence - 2 - larged trained on DOTA - v1.0 with LMMRotate. More checkpoints for aerial detection with LMMRotate in our paper can be found in this repo.

👥 Authors

💡 Usage Tip

If you find our work helpful, please consider giving us a ⭐!

🔗 Links

📦 Installation

Downloading Guide

You can download with your web browser on the file page.

We recommend downloading in the terminal using huggingface-cli (pip install --upgrade huggingface_cli). You can refer to the document for more usages.

# Set Huggingface Mirror for Chinese users (if required):
export HF_ENDPOINT=https://hf-mirror.com 
# Download a certain checkpoint:
huggingface-cli download Qingyun/Florence-2-large-DOTA-v1.0-lmmrotate --repo-type model --local-dir checkpoint/Florence-2-large-DOTA-v1.0-lmmrotate/
# If any error (such as network error) interrupts the downloading, you just need to execute the same command, the latest huggingface_hub will resume downloading.

📚 Documentation

Detection Performance

📄 License

This project is licensed under the MIT license.

📖 Cite

Please cite the LMMRotate paper if you use this code in your research:

@article{li2025lmmrotate,
  title={A Simple Aerial Detection Baseline of Multimodal Language Models},
  author={Li, Qingyun and Chen, Yushi and Shu, Xinya and Chen, Dong and He, Xin and Yu Yi and Yang, Xue },
  journal={arXiv preprint arXiv:2501.09720},
  year={2025}
}

Property	Details
Model Type	microsoft/Florence - 2 - large
Training Data	Qingyun/lmmrotate - sft - data
Pipeline Tag	image - text - to - text
Tags	aerial, geoscience, remotesensing

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご