How I ended up maintaining a python package with over 350,000+ downloads

Kurian Benoy (~kurianbenoy)


2

Votes

Description:

This talk is a sequel of Kurian's previous Pycon India 2023 talk OpenAI Whisper and it's amazing power to do fine-tuing in my mother tongue. For that talk one of the core-part was how to fine-tune and my python-library for benchmarking ASR for Malayalam languages - malayalam-asr-benchmarking. When building that library, Kurian ended up building another python-library named whisper_normalizer.

The main reason for that, I felt details about the text normalization approach used by whisper which can be found on Appendix Section C pp.21 the paper Robust Speech Recognition via Large-Scale Weak Supervision by OpenAI team is super useful. In my work, I had written another internal library and I was using same stuff again for this project as well. I was super familar with nbdev framework and releasing it was not that hard for me. TBH in my previous talk, I didn't even have one slide about this package. Looking back, now it looks like a good problem to have.

Fast track to December 2023, I noticed this github project all of a sudden has like 30+ stars. It was surprising to me and before thinking too much I realized that I got some months like 50K+ downlads. Then downloads started increasing, and it's constantly increasing all the time. At the time of writing this proposal the number of downloads for my package is as shown in the tweet.

Talk outline:

TBD

Slides: TBD

Prerequisites:

  • No pre-requisites.

Speaker Info:

I am working as an ML Engineer at Speech Team building Full-Stack GenAI @ Sarvam.ai. I have multiple years of experience in Python and Machine learning experience. I have contributed to various open source organizations like Swathanthra Malayalam Computing, FOSSASIA, Keras, DVC, HuggingFace, fast.ai, CloudCV etc.

I am the maintainer of whisper_normalizer python package which has 230,000+ downloads so far and various Malayalam ASR models like MalWhisper and VegamWhisper. I am a core-developer of indic-subtitler website as well, which is an open source subtitling platform for transcribing and translating videos/audios in Indic languages.

More details about my previous talks can be found by clicking the link.

Section: Artificial Intelligence and Machine Learning
Type: Talk
Target Audience: Intermediate
Last Updated: