LF Deep Learning Foundation today announced the first publicly available release of Acumos AI, an open source framework and platform for the training and deployment of AI models.
Created in March, the LF Deep Learning Foundation is part of the Linux Foundation project and supports open source projects for machine learning, deep learning, and AI.
Founding members include Tencent, Baidu, Huawei, ZTE, AT&T, and Nokia.
Acumos AI, whose release version is codenamed Athena, also began in March and includes the participation of about 75 developers, a foundation spokesperson told VentureBeat in an email. An updated version is due out in mid-2019, according to a statement provided to VentureBeat.
The first Acumos AI public release can deliver one-click deployment of AI models with Docker or Kubernetes containers, data translation tools, and a graphical user interface for linking multiple AI models together.
Acumos works with AI models created with tools like TensorFlow and scikit-learn, as well as H20.ai.
Other active LF Deep Learning Foundation projects include machine learning platform Angel and Elastic Deep Learning, a project to help cloud service providers make cloud cluster services with frameworks like TensorFlow and PaddlePaddle.
The two projects were added in August by Tencent and Baidu, respectively, a foundation spokesperson told VentureBeat in an email.
The Acumos AI community will also be able to utilize the AI Marketplace for sharing and downloading models. To seed the marketplace, a $100,000 competition was held earlier this year, with top winners making models that do things like predict the price of a home or determine the malignancy of a strain of breast cancer.
Acumos is the latest tool for quick AI deployment to make its debut in recent weeks. There’s also Horizon, a reinforcement learning platform created by Facebook engineers and researchers, as well as Kubeflow Pipelines from Google Cloud Platform, which also relies on containers for deploying AI.
The audio problem: Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here