提交 c47aebf6 编写于 作者: M MaoXianxin

后向重计算在OneFlow中的实现:以时间换空间,大幅降低显存占用

上级 dbdf1c08
......@@ -363,7 +363,7 @@ https://github.com/Oneflow-Inc/OneFlow-Benchmark/tree/master/LanguageModeling/GP
注:题图源自insspirito,pixabay
参考文献
参考文献
[1] Tianqi Chen, Bing Xu, Chiyuan Zhang, and Carlos Guestrin. Training Deep Nets with Sublinear Memory Cost. arXiv preprint arXiv:1604.06174, 2016.
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册