From 26e14378e82e567182909507c48a4e7fdd185f3b Mon Sep 17 00:00:00 2001 From: skirui-source Date: Mon, 17 Jul 2023 11:09:15 -0700 Subject: [PATCH] few edits --- source/examples/xgboost-rf-gpu-cpu-benchmark/notebook.ipynb | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/source/examples/xgboost-rf-gpu-cpu-benchmark/notebook.ipynb b/source/examples/xgboost-rf-gpu-cpu-benchmark/notebook.ipynb index 7e6ef965..83ea20d6 100644 --- a/source/examples/xgboost-rf-gpu-cpu-benchmark/notebook.ipynb +++ b/source/examples/xgboost-rf-gpu-cpu-benchmark/notebook.ipynb @@ -13,7 +13,7 @@ "\n", "For instance, we find a x speedup in wall clock time (6 hours vs 3+ days) and a x reduction in cost when comparing between GPU and CPU EC2 instances on 100 XGBoost HPO trials using 10 parallel workers on 10 years of the Airline Dataset.\n", "\n", - "For more check out our AWS blog(link)." + "For more check out our blog(link)." ] }, { @@ -102,9 +102,9 @@ "id": "533be0b1-0d5e-46b3-9ff1-dd71751fe68f", "metadata": {}, "source": [ - "To maximize on efficiency, we launch a `LocalCUDACluster` that utilizes GPUs for distributed computing. Then connect a Dask Client to submit and manage computations on the cluster. Refer to this (link) for more information on how to achieve this.\n", + "To maximize on efficiency, we launch either a `LocalCluster` for cpu or `LocalCUDACluster` that utilizes GPUs for distributed computing. Then connect a Dask Client to submit and manage computations on the cluster. \n", "\n", - "Submit dataset to the Dask client, instructing Dask to store the dataset in memory at all times. This can improve performance by avoiding unnecessary data transfers during the hpo process. \n" + "Submit and \"persist\" the dataset to the Dask client, which instructs Dask to store the dataset in memory at all times for faster performance by avoiding unnecessary data transfers during the hpo process. \n" ] }, {