lxj321/Multi-config-Radiomap-Dataset

Multi-config Radiomap Dataset and Pretrained Models for U6G XL-MIMO

This repository provides the public release of the Multi-config Radiomap Dataset and pretrained models for U6G / XL-MIMO radiomap prediction.

It includes:

• a large-scale radiomap dataset across 800 urban scenes
• multiple frequency bands and array configurations
• beam-map-related benchmark resources
• pretrained models for benchmark tasks

Links

• Paper: https://arxiv.org/abs/2603.06401
• Project Website: https://lxj321.github.io/MulticonfigRadiomapDataset/
• Code Repository: https://github.com/Lxj321/MulticonfigRadiomapDataset
• Dataset + Pretrained Models: this Hugging Face repository

Contents

Files in this repository

• Dataset_*.zip
  Main dataset package, including radiomap-related data and associated resources.

• Pretrained_Model_*.zip
  Pretrained models for benchmark tasks.

• metadata.csv
  Lightweight metadata index for preview and quick inspection.

Dataset Summary

This project is designed for studying:

• multi-configuration radiomap prediction
• cross-configuration generalization
• cross-environment generalization
• beam-aware radiomap modeling
• sparse radiomap reconstruction

Quick facts

• Scenes: 800
• Frequency bands: 1.8 / 2.6 / 3.5 / 4.9 / 6.7 GHz
• TX antenna scale: up to 32x32 UPA
• Beam settings: 1 / 8 / 16 / 64 beams

Visual Examples

Representative examples from the released dataset are shown below.

Overview

This figure shows an overall preview of the released data components, including the height map, ray-tracing radiomap, and configuration-only beam map.

Radiomap Examples

Representative ray-tracing radiomap examples from different scenes and configurations.

Height Map Examples

Representative height map examples from the released urban scenes.

Beam Map Examples

Representative configuration-only beam map examples.

Paired Examples

The following examples illustrate the correspondence among the released data components for selected scenes and transmitter configurations.

Example 1

Example 2

Example 3

Example 4

Example 5

Intended Usage

This dataset is intended for:

• benchmark evaluation of radiomap prediction methods
• studying generalization across unseen array configurations
• studying generalization across unseen environments
• evaluating physics-informed features such as beam maps
• reproducing the results of the associated benchmark project

Download and Usage

Download the released zip packages from the Files and versions tab.

For code, preprocessing, training, evaluation, and benchmark usage, please refer to:

• GitHub: https://github.com/Lxj321/MulticonfigRadiomapDataset
• Project Website: https://lxj321.github.io/MulticonfigRadiomapDataset/

Repository Structure

The released resources are organized around:

• dataset files
• pretrained model files
• project documentation
• benchmark code in the GitHub repository

Download and Usage

Download the released zip packages from the Files and versions tab.

For code, preprocessing, training, evaluation, and benchmark usage, please refer to:

• GitHub: https://github.com/Lxj321/MulticonfigRadiomapDataset
• Project Website: https://lxj321.github.io/MulticonfigRadiomapDataset/

Repository Structure

The released resources are organized around:

• dataset files
• pretrained model files
• project documentation
• benchmark code in the GitHub repository

Citation

If you use this dataset or the pretrained models, please cite the associated project and paper.

@misc{li2026u6gxlmimoradiomapprediction,
      title={U6G XL-MIMO Radiomap Prediction: Multi-Config Dataset and Beam Map Approach}, 
      author={Xiaojie Li and Yu Han and Zhizheng Lu and Shi Jin and Chao-Kai Wen},
      year={2026},
      eprint={2603.06401},
      archivePrefix={arXiv},
      primaryClass={eess.SP},
      url={https://arxiv.org/abs/2603.06401}, 
}

Formal citation information will be updated after the paper metadata is finalized.

License

• Dataset: CC BY 4.0
• Code: see the GitHub repository license
• Pretrained models: released together with this dataset repository unless otherwise specified

Contact

Xiaojie Li xiaojieli@seu.edu.cn xiaojieli@nuaa.edu.cn