[英]Where is tf.contrib.layers.recompute_grad in TensorFlow 2.0?
Does TensorFlow 2.0 have any support for making a computation as "recompute during the gradient pass" to save memory? TensorFlow 2.0是否支持将计算作为“渐变过程中的重新计算”以节省内存? TensorFlow 1.x has tf.contrib.layers.recompute_grad , but contrib
is gone in TF 2.0 and it doesn't look like anyone moved recompute_grad
. TensorFlow 1.x具有tf.contrib.layers.recompute_grad ,但是contrib
在TF 2.0中已经消失了,看起来好像没有人搬过 recompute_grad
。
Ya, they didn't move it. 是的,他们没有移动它。 This is a big error on Google's part. 这是Google的重大错误。 I won't move to Tensorflow 2 for this reason. 因此,我不会移至Tensorflow 2。 Recomputing gradients and creating memory efficient graphs is EXTREMELY helpful, if not vital. 重新计算梯度并创建内存有效的图形非常有用,即使不是很重要。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.