wi-lab/lwm · Dataset files are corrupted

Hello,
The DeepMIMO scenario files uploaded here do not work with the DeepMIMO Python package. You can reproduce this error by cloning this repository and trying to generate channels with DeepMIMO or by trying to open the .mat files in MATLAB. However, downloading the same city scenario folder from the DeepMIMO website fixes the issue.

Unfortunately, a similar problem occurs with the .p files you shared for downstream tasks. These files cannot be run in MATLAB. I suppose you meant these files to be used to generate the train and test sets for both downstream tasks, but this is not possible currently.

I believe it would be very beneficial for the research community if you could ensure your results are reproducible.

Hello, thank you for your feedback!

Please try cloning the scenarios again. If any issues persist, you may need to manually download them from the dataset repository.
The .p files associated with the challenge problems on the LWM website—LoS/NLoS Classification and Beam Prediction—are Torch tensors. To convert them into MATLAB-compatible .mat files, please use the p2mat.py script available at:
p2mat.py on Hugging Face.

For the downstream tasks in the paper, we used the following Sub-6 GHz channel scenarios:

city_18_denver
city_15_indianapolis
city_19_oklahoma
city_12_fortworth
city_11_santaclara
city_7_sandiego

Their mmWave counterparts, which were used for generating the best beam labels, can be found at:
mmWave Scenarios on Hugging Face.

Please let us know if you have any questions!