TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Label Studio v1.0

34 点作者 deppp大约 4 年前

2 条评论

deppp大约 4 年前
Excited to post about this, Label Studio – open source data labeling tool we&#x27;ve started working on more than a year ago is hitting v1.0 with a lot of new goodies.<p>We originally started working on it due to the complexity of bringing ML systems to production. So much depends on data quality, and dataset understanding, but the tooling to help you with that process was just not there. You probably know that first hand if you&#x27;ve worked on production ML systems.<p>With this release Label Studio gets multi-user labeling. It&#x27;s important because advanced and accurate supervised ML systems are a product of collaboration between people. Using accounts users can work on the same datasets and each user’s annotations are tied to their account. Using Projects you can streamline managing and work on different datasets, projects as well can be shared with other users. It can be embedded into the ML pipeline for humans in the loop and online learning scenarios.You can use Label Studio for computer vision, NLP, speech processing, time series analysis, and more<p>Would love to hear your feedback!
hootyhoo大约 4 年前
<a href="https:&#x2F;&#x2F;medium.com&#x2F;syncedreview&#x2F;data-annotation-the-billion-dollar-business-behind-ai-breakthroughs-d929b0a50d23" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;syncedreview&#x2F;data-annotation-the-billion-...</a>