From 71f79cb522045e1d928a7c9c9daf0209621c39f8 Mon Sep 17 00:00:00 2001 From: r-sarma Date: Thu, 2 May 2024 18:45:16 +0200 Subject: [PATCH] Add README --- .../tf-tutorial-1-imagenet/README.md | 20 +++++++++++++++++++ 1 file changed, 20 insertions(+) create mode 100644 tutorials/distributed-ml/tf-tutorial-1-imagenet/README.md diff --git a/tutorials/distributed-ml/tf-tutorial-1-imagenet/README.md b/tutorials/distributed-ml/tf-tutorial-1-imagenet/README.md new file mode 100644 index 00000000..c2c49595 --- /dev/null +++ b/tutorials/distributed-ml/tf-tutorial-1-imagenet/README.md @@ -0,0 +1,20 @@ +# Tutorial: distributed strategies for Tensorflow + +In this tutorial we show how to use Tensorflow `MultiWorkerMirroredStrategy`. +Note that the environment is tested on the HDFML system at JSC. +For other systems, the module versions might need change accordingly. +Other strategies will be updated here. + +First, from the root of this repository, build the environment containing +Tensorflow. You can *try* with: + +```bash +# Creates a Python venv called envAItf_hdfml +make tf-gpu-jsc +``` + +If you want to distribute the code in `train.py`, run from terminal: + +```bash +sbatch tfmirrored_slurm.sh +```