Well-funded Hertex’s machine learning data label startup opens up new use cases on the platform today, making it much easier to annotate audio files across commercial and open-source versions of Label Studio. For the users. These updates include a new user interface optimized for working with audio files, the ability to annotate long files (up to two hours long), and millisecond controls in the built-in annotation tool. There is now a new rendering engine for displaying waveforms. In the previous update, Heartex also added frame-by-frame video object tracking and a stripped down annotation UI to its device features.
The company says that around 150,000 users currently rely on Label Studio, with more than 95 million annotations created so far.
“When we were initially thinking about building a data identification solution, we interviewed a lot of data scientists,” Michael Maluk, founder and CEO of Heartex, told me. “And what we’ve found is that many say they don’t have enough flexibility with existing tools. There is a tool for images, but you can only put a single image on the screen. There’s a tool for our voice, but it’s very hard code from a use case perspective. And for Account Studio, when data scientists have a question mark in their head, ‘Does it support my use case?’ The answer should always be yes – it should always support your use case.
And, of course, to name data, all you need is a subject matter expert, not a software engineer. The idea behind Tag Studio is to enable almost anyone to tag your data. “We think every AI company is going to turn into a data analytics company or a data set development company,” Maluk said. Basically, Heartex says it wants to make Label Studio the primary IDE for database development.
Looking ahead, the company plans to invest heavily in the user community and plans to host its first Label Studio user conference next year.
Leave a Reply