Raw data of "Leveraging network motifs to improve artificial neural networks"

This dataset is associated with the publication "Leveraging network motifs to improve artificial neural networks". It contains a total of 188 .npy, 33 .pkl, and 8 .avi files, occupying approximately 91.4 GB of storage. The 188 ".npy" and 33 ".pkl" files were generated using the script "run_1_tasks.py" across five experiments, while the 8 ".avi" files were produced using the script "show_video.py".
The project has been under continuous development for over 1,800 days. To reproduce all experiments and obtain consistent or similar (due to randomization) output files using an 11th Gen Intel(R) Core(TM) i7-11370H @ 3.30 GHz, one may require several months of computation. Therefore, as a pilot project in collaboration with Figshare, we have stored all intermediate and final data associated with the experimental or analytical process(es). This enables readers to verify results starting from any stage or between any two stages of the processes, as well as to conduct further analyses based on these intermediate data.
The following are basic recommendations for using this dataset: It is recommended to first consult the corresponding GitHub repository to identify which portions of the raw data are required for your specific purpose. After determining the relevant subset, download only those raw data files and place them in the designated folder. You may then run "run_1_tasks.py" to reproduce the corresponding results, or modify the code to achieve your intended objectives.