Collabora is headed to San Jose, California, to take part in the inaugural edition of AI​.dev: Open Source GenAI & ML Summit, a new event which aims to bring together the brightest developers from around the world to shape the trajectory of open source AI.

Join us on Tuesday, December 12, as Jakub Piotr Cłapa dives into findings from WhisperSpeech, a new Open Source text-to-speech model developed by Collabora. Based entirely on properly licensed speech datasets and unrestricted Open Source code, the model's focus is to deliver the best natural-sounding Open Source speech synthesis solution for improved communication.

In this talk, Jakub will look at how Collabora scaled its models and training pipelines from hundreds to 80K+ hours of speech recordings, and will share lessons learned along the way. He'll also discuss some of the challenges encountered, including:

• Gone in 16 minutes: the importance of small scale experiments.
• Full throttle: is 100% GPU utilization enough?
• Do you need a fancy framework? From single- to multi-GPU training.
• Are SSDs fast enough? WebDataset brings a 10x improvement.
• Does bigger always mean better? How to effortlessly scale AI models.
• Clouds, enthusiasts or clusters? How to hunt down GPUs.
• Defending moats. How is a gaming 4090 different from an H100?

If you plan on attending, please make sure to come say hello! Note that so you can also watch Jakub's talk remotely via the Room LL20D live stream.

Update: The video recording is now available, click on the link below to start watching!

Collabora @ AI​.dev: Open Source GenAI & ML Summit

Tricks Learned from Scaling WhisperSpeech Models to 80k+ Hours of Speech
Presented by Jakub Piotr Cłapa - Tuesday, December 12

Related Posts

WhisperSpeech: Exploring new horizons in text-to-speech tech

Benchmarking machine learning frameworks

MLfix to quickly fix datasets

Related Posts

WhisperSpeech: Exploring new horizons in text-to-speech tech

Benchmarking machine learning frameworks

MLfix to quickly fix datasets