There are two Jupiter Notebooks notes in the directory

The first file(Commands Recognition) discloses reveals the process of teaching the command recognition model.

https://www.kaggle.com/code/dimalevch/commands-recognition

The second file(Transformer) contains a code that allows you to train, download and predict the transcription of the sound wave, where the speech is present. Training took place over a pre-trained transformer model.

P.S. Thanks to Kevin(Kaggle - kevinvaishnav) for participating in the project, teaching this model and providing the necessary resources!

https://www.kaggle.com/code/kevinvaishnav/stt-readymade/notebook