John Lockman 9fb395a4f5 Delete k8s-TensorFlow-resnet50-multinode-MPIOperator.yaml | %!s(int64=2) %!d(string=hai) anos | |
---|---|---|
.. | ||
PyTorch | %!s(int64=2) %!d(string=hai) anos | |
TensorRT-InferenceServer | %!s(int64=4) %!d(string=hai) anos | |
login_node_example | %!s(int64=3) %!d(string=hai) anos | |
README.md | %!s(int64=4) %!d(string=hai) anos | |
device_ip_list.yml | %!s(int64=2) %!d(string=hai) anos | |
host_inventory_file | %!s(int64=3) %!d(string=hai) anos | |
host_inventory_file.ini | %!s(int64=3) %!d(string=hai) anos | |
host_mapping_file_one_touch.csv | %!s(int64=3) %!d(string=hai) anos | |
host_mapping_file_os_provisioning.csv | %!s(int64=3) %!d(string=hai) anos | |
k8s-tensorflow-nvidia-ngc-resnet50-multinode-mpioperator.yaml | %!s(int64=4) %!d(string=hai) anos | |
mapping_device_file.csv | %!s(int64=3) %!d(string=hai) anos | |
slurm-TensorFlow-resnet50-multinode-MPI.batch | %!s(int64=4) %!d(string=hai) anos |
The examples K8s Submit and SLURM submit are provide as examples for running the resnet50 benchmark with TensorFlow on 8 GPUs using 2 C4140s.
kubectl create -f k8s-TensorFlow-resnet50-multinode-MPIOperator.yaml
sbatch slurm-TensorFlow-resnet50-multinode-MPI.batch