Modulation Recognition

Link to challenge here

Files

• parse_meta.py : retrieve general information within the project SigMF meta data
• bw_duration.py : MIN/MAX values for BW or sample count
• snr.py : compute power information on each modulation/file (linear and log power, variance and SNR)
• imgtf.py : create the Freq/Time image of a signal and the label image
• imgtf_wo_mod.py : equivalent of imagtf but without any modulation distinction - only noise and signal
• open_imag_and_label TF_image: open simultaneously the TF_image and its label equivalent using feh

Instructions

cd /app/mod_reco/deeplab2
python3 trainer/train.py --config_file="/app/mod_reco/deeplab2/configs/Wireless/resnet50_os32_semseg.textproto" --mode=train --model_dir="/app/data/Modulation_recognition/models/" --num_gpus=2 >> /app/data/Modulation_recognition/log006.log 

%cd /app/mod_reco/deeplab2
!python3 trainer/train.py --config_file="/app/mod_reco/deeplab2/configs/Wireless/resnet50_os32_semseg.textproto" --mode=eval --model_dir="/app/data/Modulation_recognition/models/" --num_gpus=1 >> /app/data/Modulation_recognition/log006_eval.log