Speech Recognition And Summarization System In Python [Project Tutorial]

We'll build a system that can recognize speech in audio files, generate a transcript, and summarize the transcript.

By the end, you'll have a full speech recognition system that will run on your computer, and be able to transcribe and summarize podcasts, lecture notes, and meeting recordings.

All of the code will be written in JupyterLab using Python. You can read an overview of the project, and see the full code, here - https://github.com/dataquestio/project-walkthroughs/tree/master/speech_recognition .

We'll start out by build a speech recognition system using a python package called vosk for the recognition and pydub for loading the audio. Then, we'll use recasepunc to add punctutation to our transcripts. We'll then use a huggingface pipeline to summarize the transcripts.

You'll need to install some packages and download 2 audio files, which is covered in the README - https://github.com/dataquestio/project-walkthroughs/blob/master/speech_recognition/README.md .

We have some viewer Q&A at the end, which may help you if you have any issues with the project.

Chapters:

00:00 Introduction
01:53 - Speech recognition using vosk
11:34 - Adding punctuation to our transcript with recasepunc
17:01 - A function to transcribe longer audio files
21:13 - Summarizing the transcripts using huggingface transformers
27:21 - Wrapping up with a project overview
28:22 - Q&A: Speech recognition using vosk
31:45 - Q&A: Adding punctuation to our transcript with recasepunc
35:03 - Q&A: A function to transcribe longer audio files
37:38 - Q&A: Summarizing the transcripts using huggingface transformers
39:26 - Q&A: Project overview

-----------------------------
Join 1M+ Dataquest learners today!
Master data skills and change your life.
Sign up for free: https://bit.ly/3O8MDef