Setup TensorFlow

TensorFlow is a widely-used open-source deep learning framework, recognized for its static computation graph that optimizes performance and scalability, making it a favored choice for deploying machine learning models at scale in various industries.

With an Ansible script, deploy TensorFlow on both kube_nodes and the kube_control_node. After the deployment of TensorFlow, you gain access to the TensorFlow container.

Prerequisites

  • Ensure nerdctl is available on all cluster nodes.

  • If GPUs are present on the target nodes, install NVidia CUDA (with containerd) or AMD ROCm drivers during provisioning. CPUs do not require any additional drivers.

  • Use local_repo.yml to create an offline TensorFlow repository. For more information, click here.

[Optional prerequisites]

  • Ensure the system has enough space.

  • Ensure the passed inventory file includes a kube_control_plane and a kube_node_group listing all cluster nodes. Click here for a sample file.

  • Nerdctl does not support mounting directories as devices because it is not a feature of containerd (The runtime that nerdctl uses). Individual files need to be attached while running nerdctl.

  • Container Network Interface should be enabled with nerdctl.

Deploying TensorFlow

  1. Change directories to the tools folder:

    cd tools

  2. Run the tensorflow.yml playbook:

    ansible-playbook tensorflow.yml -i inventory

Accessing TensorFlow (CPU)

  1. Verify that the tensorflow image present in container engine images:

    nerdctl images

  2. Use the container image per your needs:

    nerdctl run -it --rm tensorflow/tensorflow

For more information, click here.

Accessing TensorFlow (AMD)

  1. Verify that the tensorflow image present in container engine images:

    nerdctl images

  2. Use the container image per your needs:

    nerdctl run -it --network=host --device=/dev/kfd --device /dev/dri/card0 --device /dev/dri/card1 --device /dev/dri/card2 --device /dev/dri/renderD128 --device /dev/dri/renderD129  --ipc=host --shm-size 16G --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined rocm/tensorflow:latest

For more information, click here.

Accessing TensorFlow (NVidia)

  1. Verify that the tensorflow image present in container engine images:

    nerdctl images

  2. Use the container image per your needs:

    nerdctl run --gpus all -it --rm nvcr.io/nvidia/tensorflow:23.12-tf2-py3

For more information, click here.

If you have any feedback about Omnia documentation, please reach out at omnia.readme@dell.com.