Choosing a container image

This page helps you to choose which container image you want to use.

Choose a container image type

Each container image provides a Python 3 environment and includes the selected data science framework (such as PyTorch or TensorFlow), Conda, the NVIDIA stack for GPU images (CUDA, cuDNN, NCCL2), and many other supporting packages and tools. To find the container image that you want, see the table below.

The following list of AI Platform Deep Learning Containers image types is organized by framework type.

Framework	Processor	Container Image Name(s)
Base	GPU	gcr.io/deeplearning-platform-release/base-cu100 gcr.io/deeplearning-platform-release/base-cu101 gcr.io/deeplearning-platform-release/base-cu110
	CPU	gcr.io/deeplearning-platform-release/base-cpu
TensorFlow Enterprise 2.x	GPU	gcr.io/deeplearning-platform-release/tf2-gpu.2-1 gcr.io/deeplearning-platform-release/tf2-gpu.2-3
	CPU	gcr.io/deeplearning-platform-release/tf2-cpu.2-1 gcr.io/deeplearning-platform-release/tf2-cpu.2-3
TensorFlow Enterprise 1.x	GPU	gcr.io/deeplearning-platform-release/tf-gpu gcr.io/deeplearning-platform-release/tf-gpu.1-15
	CPU	gcr.io/deeplearning-platform-release/tf-cpu gcr.io/deeplearning-platform-release/tf-cpu.1-15
TensorFlow 2.x	GPU	gcr.io/deeplearning-platform-release/tf2-gpu gcr.io/deeplearning-platform-release/tf2-gpu.2-0 gcr.io/deeplearning-platform-release/tf2-gpu.2-2 gcr.io/deeplearning-platform-release/tf2-gpu.2-4
	CPU	gcr.io/deeplearning-platform-release/tf2-cpu gcr.io/deeplearning-platform-release/tf2-cpu.2-0 gcr.io/deeplearning-platform-release/tf2-cpu.2-2 gcr.io/deeplearning-platform-release/tf2-cpu.2-4
TensorFlow 1.x	GPU	gcr.io/deeplearning-platform-release/tf-gpu.1-13 gcr.io/deeplearning-platform-release/tf-gpu.1-14
	CPU	gcr.io/deeplearning-platform-release/tf-cpu.1-13 gcr.io/deeplearning-platform-release/tf-cpu.1-14
PyTorch	GPU	gcr.io/deeplearning-platform-release/pytorch-gpu gcr.io/deeplearning-platform-release/pytorch-gpu.1-0 gcr.io/deeplearning-platform-release/pytorch-gpu.1-1 gcr.io/deeplearning-platform-release/pytorch-gpu.1-2 gcr.io/deeplearning-platform-release/pytorch-gpu.1-3 gcr.io/deeplearning-platform-release/pytorch-gpu.1-4 gcr.io/deeplearning-platform-release/pytorch-gpu.1-6 gcr.io/deeplearning-platform-release/pytorch-gpu.1-7 gcr.io/deeplearning-platform-release/pytorch-gpu.1-8
	CPU	gcr.io/deeplearning-platform-release/pytorch-cpu gcr.io/deeplearning-platform-release/pytorch-cpu.1-0 gcr.io/deeplearning-platform-release/pytorch-cpu.1-1 gcr.io/deeplearning-platform-release/pytorch-cpu.1-2 gcr.io/deeplearning-platform-release/pytorch-cpu.1-3 gcr.io/deeplearning-platform-release/pytorch-cpu.1-4
PyTorch XLA	TPU/GPU/CPU (experimental)	gcr.io/deeplearning-platform-release/pytorch-xla.1-6 gcr.io/deeplearning-platform-release/pytorch-xla.1-7 gcr.io/deeplearning-platform-release/pytorch-xla.1-8
R	CPU (experimental)	gcr.io/deeplearning-platform-release/r-cpu gcr.io/deeplearning-platform-release/r-cpu.3-6
Scikit-learn	CPU (experimental)	gcr.io/deeplearning-platform-release/sklearn-cpu gcr.io/deeplearning-platform-release/sklearn-cpu.0-23
XGBoost	CPU (experimental)	gcr.io/deeplearning-platform-release/xgboost-cpu gcr.io/deeplearning-platform-release/xgboost-cpu.1-1

Note: Deep Learning Containers that do not include a version number (for example, pytorch-gpu) refer to the latest release version of that framework. If your workload requires a specific framework version, specify that version.

TensorFlow Enterprise container images

TensorFlow Enterprise container images provide you with a Google Cloud optimized distribution of TensorFlow with Long Term Version Support. To learn more about TensorFlow Enterprise, read the TensorFlow Enterprise overview.

Experimental images

Some Deep Learning Containers image families are experimental, as indicated by the table of image families. Experimental images are supported on a best-effort basis, and may not receive refreshes on each new release of the framework.

Listing all available versions

If you need a specific framework or CUDA version, you can search the complete list of available container images. To list all available Deep Learning Containers images, use the following command in the gcloud command-line tool with your preferred terminal or in Cloud Shell.

gcloud container images list --repository="gcr.io/deeplearning-platform-release"

Using locally

Deep Learning Containers can be pulled and used locally. To do so, see Getting started with a local deep learning container.

What's next

• Read the Deep Learning Containers overview to learn more about what is pre-installed on container images.
• Get started with Deep Learning Containers by walking through the How-to guides, which provide instructions on how to build and push deep learning container images.