iRadioDiff

---

Welcome to the RadioDiff family

Base BackBone, Paper Link: RadioDiff, Code Link: GitHub

PINN Enhanced with Helmholtz Equation, Paper Link: RadioDiff-$k^2$, Code Link: GitHub

Efficiency Enhanced RadioDiff, Paper Link: RadioDiff-Turbo

Indoor RM Construction with Physical Information, Paper Link: iRadioDiff, Code Link: GitHub

3D RM with DataSet, Paper Link: RadioDiff-3D, Code Link: GitHub

Sparse Measurement for RM ISAC, Paper Link: RadioDiff-Inverse

Sparse Measurement for NLoS Localization, Paper Link: RadioDiff-Loc

For more RM information, please visit the repo of Awesome-Radio-Map-Categorized

---

This is the code of "iRadioDiff: Physics Informed Diffusion Model for Effective Indoor Radio Map Construction and Localization" accepted by the IEEE ICC 2026.

☀️ Before Starting

1. install torch

conda create -n radiodiff python=3.9
conda avtivate radiodiff
pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113

2. install other packages.

pip install -r requirement.txt

3. prepare accelerate config.

accelerate config # HOW MANY GPUs YOU WANG TO USE.

🎇 Prepare Data

We used the Indoor Radio Map Dataset dataset for model training and testing.

• The data structure should look like:

|-- $ICASSP2025_Dataset
|   |-- Input
|   |-- |-- Task_1_ICASSP
|   |-- |-- |-- B1_Ant1_f1_S0.PNG
|   |-- |-- |-- B1_Ant1_f1_S1.PNG
|   ...
|   |-- Output
|   |-- |-- Task_1_ICASSP
|   |-- |-- |-- B1_Ant1_f1_S0.PNG
|   |-- |-- |-- B1_Ant1_f1_S1.PNG
|	...

🎉 Training

accelerate launch train_cond_ldm.py --cfg ./configs/ICA_dm.yaml

V. Inference.

make sure your model weight path is added in the config file ./configs/ICA_dm.yaml (line 66), and run:

python sample_cond_ldm.py --cfg ./configs/ICA_dm.yaml

Note that you can modify the sampling_timesteps (line 7) to control the inference speed.

Thanks

Thanks to the base code DDM-Public.