Installation

DALI and NGC

DALI is preinstalled in the NVIDIA GPU Cloud TensorFlow, PyTorch, and MXNet containers in versions 18.07 and later.

Installing prebuilt DALI packages

Prerequisites

  1. Linux x64.

  2. NVIDIA Driver supporting CUDA 10.0 or later (i.e., 410.48 or later driver releases).

  3. One or more of the following deep learning frameworks:

Installation

Execute the following command to install latest DALI for specified CUDA version:

  • for CUDA 10:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist nvidia-dali-cuda100

  • for CUDA 11:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist nvidia-dali-cuda110

DALI TensorFlow plugin (nvidia-dali-tf-plugin)

DALI doesn’t contain prebuilt versions of the DALI TensorFlow plugin. It needs to be installed as a separate package which will be built against the currently installed version of TensorFlow:

  • for CUDA 10:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist nvidia-dali-tf-plugin-cuda100

  • for CUDA 11:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist nvidia-dali-tf-plugin-cuda110

Installing this package will install nvidia-dali-cudaXXX and its dependencies, if they are not already installed. The package tensorflow-gpu must be installed before attempting to install nvidia-dali-tf-plugin-cudaXXX.

Note

The packages nvidia-dali-tf-plugin-cudaXXX and nvidia-dali-cudaXXX should be in exactly the same version. Therefore, installing the latest nvidia-dali-tf-plugin-cudaXXX, will replace any older nvidia-dali-cudaXXX version already installed. To work with older versions of DALI, provide the version explicitly to the pip install command.

For older versions of DALI (0.22 and lower), use the package nvidia-dali. The CUDA version can be selected by changing the pip index:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/cuda/9.0 nvidia-dali
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/cuda/9.0 nvidia-dali-tf-plugin

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/cuda/10.0 nvidia-dali
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/cuda/10.0 nvidia-dali-tf-plugin

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/cuda/11.0 nvidia-dali
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/cuda/11.0 nvidia-dali-tf-plugin

CUDA 9 build is provided up to DALI 0.22.0. CUDA 10 build is provided starting from DALI 0.8.0. CUDA 11 build is provided starting from DALI 0.22.0.

Pre-built packages in Watson Machine Learing Community Edition

IBM publishes pre-built DALI packages as part of Watson Machine Learning Community Edition (WML CE). WML CE includes conda packages for both IBM Power and x86 systems. The initial release includes DALI 0.9 built against CUDA 10.1 and with TensorFlow support. Other versions may be added in the future. The WML CE conda channel also includes the CUDA prerequisites for DALI.

After installing conda and configuring the WML CE conda channel (see WML CE installation) you can install DALI:

$ conda create -y -n my-dali-env python=3.6 dali

$ conda activate my-dali-env

(my-dali-env) $ conda list dali
...
dali                      0.9             py36_666ce55_1094.g70c071f

Nightly and weekly release channels

Note

While binaries available to download from nightly and weekly builds include most recent changes available in the GitHub some functionalities may not work or provide inferior performance comparing to the official releases. Those builds are meant for the early adopters seeking for the most recent version available and being ready to boldly go where no man has gone before.

Note

It is recommended to uninstall regular DALI and TensorFlow plugin before installing nightly or weekly builds as they are installed in the same path

Nightly builds

To access most recent nightly builds please use flowing release channel:

  • for CUDA 10:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/nightly nvidia-dali-nightly-cuda100  nvidia-dali-tf-plugin-nightly-cuda100

  • for CUDA 11:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/nightly nvidia-dali-nightly-cuda110 nvidia-dali-tf-plugin-nightly-cuda110

Weekly builds

Also, there is a weekly release channel with more thorough testing. To access most recent weekly builds please use flowing release channel:

  • for CUDA 10:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/weekly nvidia-dali-weekly-cuda100 nvidia-dali-tf-plugin-weekly-cuda100

  • for CUDA 11:

pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/weekly nvidia-dali-weekly-cuda110 nvidia-dali-tf-plugin-weekly-cuda110