Google's new AI training technology JEST trains 13 times faster...
【数据猿导读】 Google's new AI training technology JEST trains 13 times faster
On July 14, it was reported that the research team of Google DeepMind recently published a breakthrough study, proposing a new method called JEST (Joint Example Selection), which aims to improve the efficiency of large-scale multimodal learning. Google claims that JEST can dramatically improve the training speed and energy efficiency of AI models by an order of magnitude through its innovative joint sample selection and learning scoring mechanism, outperforming the latest models with up to 13 times fewer iterations and 10 times fewer computations.
来源：DIYuan
声明：数据猿尊重媒体行业规范，相关内容都会注明来源与作者；转载我们原创内容时，也请务必注明“来源：数据猿”与作者名称，否则将会受到数据猿追责。
刷新相关文章
我要评论
不容错过的资讯
大家都在搜
