Skip to content
Go to file


Kaggle Notebooks allow users to run a Python Notebook in the cloud against our competitions and datasets without having to download data or set up their environment.

This repository includes our Dockerfiles for building the CPU-only and GPU image that runs Python Notebooks on Kaggle.

Our Python Docker images are stored on Google Container Registry at:

Note: The base image for the GPU image is our CPU-only image. The gpu.Dockerfile adds a few extra layers to install GPU related libraries and packages (cuda, libcudnn, pycuda etc.) and reinstall packages with specific GPU builds (torch, tensorflow and a few mores).

Getting started

To get started with this image, read our guide to using it yourself, or browse Kaggle Notebooks for ideas.

Requesting new packages

First, evaluate whether installing the package yourself in your own notebooks suits your needs. See guide.

If you the first step above doesn't work for your use case, open an issue or a pull request.

Opening a pull request

  1. Update the Dockerfile
    1. For changes specific to the GPU image, update the gpu.Dockerfile.
    2. Otherwise, update the Dockerfile.
  2. Follow the instructions below to build a new image.
  3. Add tests for your new package. See this example.
  4. Follow the instructions below to test the new image.
  5. Open a PR on this repo and you are all set!

Building a new image



  • --gpu to build an image for GPU.
  • --use-cache for faster iterative builds.

Testing a new image

A suite of tests can be found under the /tests folder. You can run the test using this command:



  • --gpu to test the GPU image.

Running the image

For the CPU-only image:

# Run the image built locally:
docker run --rm -it kaggle/python-build /bin/bash
# Run the pre-built image from
docker run --rm -it /bin/bash

For the GPU image:

# Run the image built locally:
docker run --runtime nvidia --rm -it kaggle/python-gpu-build /bin/bash
# Run the image pre-built image from
docker run --runtime nvidia --rm -it /bin/bash

To ensure your container can access the GPU, follow the instructions posted here.

Tensorflow custom pre-built wheel

A Tensorflow custom pre-built wheel is used mainly for:

  • Faster build time: Building tensorflow from sources takes ~1h. Keeping this process outside the main build allows faster iterations when working on our Dockerfiles.

Building Tensorflow from sources:

  • Increase performance: When building from sources, we can leverage CPU specific optimizations
  • Is required: Tensorflow with GPU support must be built from sources

The Dockerfile and the instructions can be found in the tensorflow-whl folder/.

You can’t perform that action at this time.