-x HOROVOD_HIERARCHICAL_ALLREDUCE=1 -x HOROVOD_FUSION_THRESHOLD=16777216 -x NCCL_MIN_NRINGS=8 -x NCCL_LAUNCH_MODE=PARALLE

dann のブックマーク 2020/11/09 00:28

<blockquote class="hatena-bookmark-comment"><a class="comment-info" href="https://b.hatena.ne.jp/entry/4693957688210134274/comment/dann" data-user-id="dann" data-entry-url="https://b.hatena.ne.jp/entry/s/aws.amazon.com/blogs/compute/scalable-deep-learning-training-using-multi-node-parallel-jobs-with-aws-batch-and-amazon-fsx-for-lustre/" data-original-href="https://aws.amazon.com/blogs/compute/scalable-deep-learning-training-using-multi-node-parallel-jobs-with-aws-batch-and-amazon-fsx-for-lustre/" data-entry-favicon="https://cdn-ak2.favicon.st-hatena.com/64?url=https%3A%2F%2Faws.amazon.com%2Fblogs%2Fcompute%2Fscalable-deep-learning-training-using-multi-node-parallel-jobs-with-aws-batch-and-amazon-fsx-for-lustre%2F" data-user-icon="/users/dann/profile.png">Scalable deep learning training using multi-node parallel jobs with AWS Batch and Amazon FSx for Lustre | Amazon Web Services</a><ul class="comment-tag" style="list-style: none; margin: 0px;"><li style="float: left">[<a href="https://b.hatena.ne.jp/q/horovod">horovod</a>]</li><li style="float: left">[<a href="https://b.hatena.ne.jp/q/mxnet">mxnet</a>]</li><li style="float: left">[<a href="https://b.hatena.ne.jp/q/aws">aws</a>]</li><li style="float: left">[<a href="https://b.hatena.ne.jp/q/tensorflow">tensorflow</a>]</li></ul><br><p style="clear: left"> -x HOROVOD_HIERARCHICAL_ALLREDUCE=1 -x HOROVOD_FUSION_THRESHOLD=16777216 -x NCCL_MIN_NRINGS=8 -x NCCL_LAUNCH_MODE=PARALLE</p><a class="datetime" href="https://b.hatena.ne.jp/dann/20201109#bookmark-4693957688210134274"><span class="datetime-body">2020/11/09 00:28</span></a></blockquote><script src="https://b.st-hatena.com/js/comment-widget.js" charset="utf-8" async></script>

このブックマークにはスターがありません。
最初のスターをつけてみよう！