YOLOR-radio

YOLOR-radio is a fine-tuned object detection model for BS identification for beam initialization to detect radio in one inference pass. The model is trained on imagery of Sivers Semiconductors 60 GHz mmWave Radio frontends (EVK06002). Part of the YOLOR detector family used for the Look Once, Beam Twice mmWave V2X beam-management pipeline (SECON 2026).

Reference implementation for the paper:

Avhishek Biswas*, Apala Pramanik*, Eylem Ekici, Mehmet C. Vuran. "Look Once, Beam Twice: Camera-Primed Real-Time Double-Directional mmWave Beam Management for Vehicular Connectivity." (*equal contribution)

arXiv: https://doi.org/10.48550/arXiv.2605.05071

Quick links

• Paper (arXiv): https://doi.org/10.48550/arXiv.2605.05071 Code and Data: https://github.com/UNL-CPN-Lab/Look-Once-Beam-Twice

Architecture	YOLOv11x, 81-class output head (COCO 80 + 1 custom)
Initialization	stock yolo11x.pt
Schedule	200 epochs, cos_lr, close_mosaic=20, lr0=0.01
Training data	IndoorCOTSDataset — 3,599 train / 449 val / 451 test
Custom classes	radio (id 80)
Released checkpoint	last.pt (the converged final model)

Usage

from huggingface_hub import hf_hub_download
from ultralytics import YOLO

weights = hf_hub_download(repo_id="cpnlab/YOLOR-radio", filename="last.pt")
model = YOLO(weights)

results = model.predict("path/to/image.jpg", conf=0.25)
results[0].show()

Class indices in the returned detections: 0–79 are the standard COCO classes; 80 is radio. The model's names dict carries the same mapping.

Intended use

• Stage-1 BS-candidate detector for the Look Once, Beam Twice detector pipeline.
• General-purpose RF-hardware-aware object detection in indoor / office scenes where both COCO objects and RF radios may appear.

Citation

@inproceedings{biswas2026look,
  title     = {Look Once, Beam Twice: Camera-Primed Real-Time Double-Directional
               mmWave Beam Management for Vehicular Connectivity},
  author    = {Biswas, Avhishek and Pramanik, Apala and Ekici, Eylem and Vuran, Mehmet C.},
  booktitle = {Proc. IEEE SECON},
  year      = {2026}
}

Paper: https://doi.org/10.48550/arXiv.2605.05071 · Code: https://github.com/UNL-CPN-Lab/Look-Once-Beam-Twice

Contact

For questions about this model or the paper, contact the corresponding authors:

• Avhishek Biswas — abiswas3@huskers.unl.edu
• Apala Pramanik — apramanik2@huskers.unl.edu

Acknowledgments

Developed at the Cyber Physical Networking (CPN) Lab, School of Computing, University of Nebraska–Lincoln, in collaboration with The Ohio State University. Thanks to Sivers Semiconductors, Ettus Research, and the open-source Ultralytics, PyTorch, and Ettus UHD communities.