Click here for details on the challenge setup
The helper functions in this starter code use SigMF and CommPy. To install these dependencies:
pip install git+https://github.com/gnuradio/SigMF.git
pip install scikit-commpy==0.6.0
(The current code and dataset depends on scikit-commpy version 0.6.0; using the latest version may introduce inconsistencies with the files generated in the dataset. Future versions of the starter code and dataset will account for the updates in this dependency.)
Other dependencies include: NumPy, Matplotlib, Tensorflow (to run the bitregression
example), tqdm (for progress bar)
Please ensure that the dataset
is saved in the folder "dataset", and retains the folder hierarchy as provided -- this allows the helper functions to correctly find and load the corresponding files.
To obtain the dataset, you may use the following commands (from within the RFChallenge starter code folder):
wget -O rfc_dataset.zip "https://www.dropbox.com/s/clh4xq7u3my3nx6/rfc_dataset.zip?dl=0"
jar -xvf rfc_dataset.zip
rm rfc_dataset.zip
(Note that jar
is used here, instead of unzip
, since the zip file is larger than 4 GB. You may also use 7z
to extract the contents.)
The python notebook notebook/Demo.ipynb
demonstrates how these helper functions may be used to load the respective sigmf files from the training and validation datasets.
The python notebook notebook/QuickStart.ipynb
provides a brief overview on helper functions and code snippet that will get you started.
Refer to the python notebook notebook/Reference_Methods.ipynb
for reference methods that you can compare against!
As some of the saved files in the example methods are large (e.g. saved models and statistics), they are not included in this Git repository.
You can get get the full contents of the example
folder using the following commands:
wget -O rfc_example.zip "https://www.dropbox.com/s/j210lhqxsbx85o4/rfc_example.zip?dl=0"
unzip -o rfc_example.zip
rm rfc_example.zip
NB: unzip -o
overwrites files in the example folder. If you have modified or added files to the example folder, back up those files before running the above commands.
This step is only required if you wish to run scripts or functions provided in the example
folder.
- Dataset (Training set) (Latest Version: Jun 15, 2021)
- Example folder from starter code (Latest Version: Sep 1, 2021)