About
Our most notable software project, ort, is used by researchers, Fortune 500s, hobbyists, and startups alike to deploy ML pipelines with maximum performance everywhere from the edge to the datacenter. Our Earshot voice activity detection model is simultaneously the fastest and most accurate VAD model we tested. It was trained with our THORN optimizer, which can accelerate AI training by up to 30% and unlocks incredible efficiency at any scale.
We're also working on a collaborative data labeling platform, dipping our toes into desktop app frameworks, and training small-but-mighty instruction-based language models perfect for domain-specific finetuning -- and it'll all be open-source!
All this work is, of course, both time- and money-intensive. Your contribution will help us dedicate more time and resources into these projects and ensure they stay free, for everyone, forever. 🟦🟩
Contribute
Become a financial contributor.
Financial Contributions
💚 Fully covers our monthly hosting expenses and gets you priority support, issue resolution, and implementation of feature requests for all project... Read more
💖 This tier lets us tinker with & train new FOSS ML models, and gets you same-day* support, issue resolution, & feature request implementation for ... Read more
pyke.io is all of us
Our contributors 19
Thank you for supporting pyke.io.
Carson M
GitHub Sponsors
$274 USD
Thomas
$200 USD
Yunho Cho
backer
$175 USD
Sef
$100 USD
zlataovce
$61 USD
Laco Skokan
$50 USD
Michele Xiloy...
backer
$30 USD
mush42
backer
$25 USD
Andre Popovitch
backer
$25 USD
Bartek
$20 USD
Noah
$20 USD
Budget
Transparent and open finances.
Credit from Andre Popovitch to pyke.io •
Credit from Michele Xiloyannis to pyke.io •
$1,098.19 USD
$1,098.19 USD
--.-- USD
$572.48 USD