/whisper_ai

Whisper AI Interview Task

Primary LanguageJupyter NotebookMIT LicenseMIT

Install requirements

This code was written using the latest release of Python 3.8.

Ensure the additional requirements are installed by running pip install -r requirements.txt in your python environment. This code should not require anything other than numpy, scipy, seaborn, matplotlib >= 3.4.0, and jupyter (for my convolution implementation).

The main submission is the file whisper_interview.py. Please note the following:

  • The functions decimate_by_2() and downsample_by_2() are defined here along with a main() function that tests the code and makes a few plots.
  • Code can be run on the command line with python whisper_interview.py to view the results of the main function call

Additionally, I have done a manual implementation of the convolution operation using just numpy and a for loop. You can see the details of that in the included jupyter notebook whisper_convolution.ipynb file.

Thank you for considering my application!

-Rohit Voleti