layer-wise adaptive rate scaling翻译adaptive
翻译为：层级自适应速率缩放
Layer-wise adaptive rate scaling is a training optimization technique used in deep learning systems. It adjusts the learning rate of different layers in the neural network as training progresses. This allows each layer to learn at an optimal rate, which may not be possible with a single, fixed learning rate.

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。