/A-Hybrid-Network-Combining-Cnn-and-Transformer-Encoder-to-Classify-Mosquitoes-Based-on-Wing-Beat-

Code for my paper : http://dx.doi.org/10.2139/ssrn.4317584

Primary LanguageJupyter NotebookMIT LicenseMIT

A-Hybrid-Network-Combining-CNN-and-Transformer-Encoder-to-Classify-Mosquitoes-Based-on-Wing-Beat-Frequencies

  1. Clone this Repo and install requirements. Go to this to look at project details.
  2. Download the Dataset. If you get 403 error just refresh the page!
  3. Extract the dataset into data folder. Dont forget to put the audio files in there respective folder names. The paths should be modified as in data.csv(sorry data is to big to upload in here:) You can refer to this as an example as its done for different duration in here.
  4. Create your profile and project name in the wandb(its free) and go to Makefile and update your information there including the api before running the makefile.
  5. run python3 run.py and your done.

or you can just play around in notebook given in the repo.