Globally distributed AI and a Constellation update
During Cloudflare's 2023 Developer Week, the company announced Constellation, a set of APIs that enable running fast, low-latency inference tasks using pre-trained machine learning/AI models directly on their network. One month after launching this feature, three new updates have been introduced: increased model size limit from 10 MB to 50 MB, tensor caching for better network latency and faster inference times, and the addition of XGBoost runtime support. Constellation allows developers to deploy globally distributed machine learning applications on Cloudflare's network.
Company
Cloudflare
Date published
June 22, 2023
Author(s)
Rita Kozlov, Celso Martinho
Word count
1255
Hacker News points
None found.
Language
English