v2.11.0

• 5d96f1f moving cuda up to 11.7
• 52807cb moving cuda to 11.5
• d3b8a7f fixing bug re: driver version
• 0f6ebe2 fixing miniconda bug in Dockerfile
• 0d9b49d moving to CUDA 11.5.0 in Dockerfile
• 3a81e30 pinning
• 69916d6 updating conda and notebook versions in the dockerfile
• 4bbb535 updating notebooks version
• f6900b2 overcoming bug in miniconda installation script
• b67a323 updating conda
• 175369b updating the environment
• 3f5d77e updating docker
• c50c357 move to driver 525
• 7cf0f32 support ubuntu 22.04

v2.9.0

• docs: Documented WSL v2 support via containers only due to CUDA 11.0 requirement in WSL v2
• fix: typos in README files
• fix: remove double quotes in cuda 11+ version reporing
• fix: checking that conda is installed before run-notebook
• fix: checking that the notebook exists in run-notebook
• fix: temporarily removing nvdashboad from create conda env pending jupyter 3.0 compatibility
• fix: quote termination in script
• fix: not including nvdashboard in Jupyter Lab pending support for Jupyter 3.0
• fix: docker file cuda version should be 11.2.2
• feat: pin (#115)
• feat: move nvdashboard to rapidsai channel (#118)
• fix: docker file cuda version should be 11.2.2
• feat: pin (#115)
• feat: move nvdashboard to rapidsai channel (#118)
• feat: remove chainer (#117)
• feat: move to RAPIDS 21.06 (#112)
• feat: move to pytorch 1.8 (#113)
• feat: remove fastai (#110)
• feat: tf2 support (#111)
• feat: notebooks 21.06 (#114)
• feat: back to conda solver(#116)
• feat: #105 update driver to 465
• feat: #106: update rapids version
• fix: dev2.9 tag
• feat: update rapids notebooks version (#107)
• feat: dev 2.9 tag (#108)
• Merge pull request #103 from hoanghphan/dev2.9
• add object_detection example
• benchmarks renamed to examples

v2.8.0

• update versions to 2.8.0
• docs: automated install on RHEL 8.x
• fix: aws cli purge must force
• docs: readme for the pytorch lightning example
• docs: added readme for the knn example
• feat: rapids cuml benchmarks
• fix: notebook version 0.18
• fix: pinning pl ver for benchmarks
• fix version comparison for driver_install
• fix: purge ngc cli
• fix: WSL v2 Ubuntu image does not have unzip
• feat:documentation (#83)
• feat: pins (#97)
• feat: bump ppa driver to 460 (#96)
• fix: aws cli
• TF2 env (#43)
• feat: AWS CLI (#65)
• feat: kaggle docker image (#69)
• feat: kaggle cli (#93)
• feat: ngc cli (#72)
• feat: notebooks 0.18 (#95)
• feat: RAPIDS 0.18
• feat: microbenchmarks (#78)
• update unsupported message for install_driver
• switch from rpmfusion.org to nvidia.com
• add install_driver() for rhel8
• fix dual detection, WSL is not mutually exclusive w/ /etc/os-release
• fix ubuntu, be consistent w/ OS_RELEASE
• fix issue where nvidia-docker relies on full distro version
• fix compound VERSION_ID issue, rhel8.3 -> rhel8
• fix redhat -> rhel to match os-release
• remove redundant sourcing of /etc/os-release
• use /etc/os-release to detect OS_FLAVOR and OS_RELEASE
• refactor changes to .bashrc in install_cuda()
• check for existence of pinned environment before trying to create a conda env from it #87
• fix 'git clone' url
• feat: upgrade the min driver and docker versions (#79)
• feat: update STACK_VERSION to 2.8.0-dev (#82)

v2.7.0

feat: update versions to 2.7.0
fix: pinning jupyterlab to 2.2.9 in the env because of nvdashboard dependency
fix: permissions on the data-science-stack script
fix: upgrade command requires bash
fix: backup old versions of envs and docker images on upgrade
fix: pin the conda envi
doc: updated readme for 2.7.0 (#73)
feat: self-upgrade (#57)
feat: support for repo2docker (#54)
feat: beta support for WSL (#28)
feat: autocomplete (#53)
feat: Proper return codes (#56)
fix: Pull 0.17 notebooks for RAPIDS 0.17
feat: 2.6.0 -> 2.7.0-dev

v2.6.0

feat: 2.6.0-dev -> 2.6.0
feat: Add adlfs, dask-glm, dask-ml, and jupyterlab-nvdashboard to default environment
feat: Respect NOTEBOOKS_DIR if set in the user's environment (#49)
feat: Update default stacks to RAPIDS 0.17
feat: Start 2.6.0-dev branch

v2.5.1

feat: 2.5.0 -> 2.5.1
docs: container data mounting, OS Support, typos

v2.5.0

Changes since v2.4.0 (latest first)

feat: 2.5.0-dev -> 2.5.0
feat: disable disk swap when installing Kubernes, and ask to turn on in purge
fix: logic around purge-driver y/n prompt always purges
feat: Kubernetes stack installer
docs: Add FAQ entry about connectivity and firewalls
feat: Command to run all notebooks in a directory
feat: Update to CUDA 11
feat: OS flavor and release logic, only run on supported OSs
fix: driver detection when driver broken
fix: Red Hat 8 docker-install, add enable and start of daemon
feat: additional stop-on-error around install commands
docs: Specify in README that Red Hat Workstation is the supported flavor
feat: After build-* commands display message about run- commands
feat: log output to data-science-stack.log
feat: Update default stacks to RAPIDS 0.16
feat: Update to 455 driver
fix: purge-driver bug on Ubuntu 20.04
feat: Start 2.5.0-dev branch

v2.4.0

Since v2.3.1 (latest first)

• feat: 2.4.0-dev -> 2.4.0
• fix: correct, explain, or reword various output and help messages
• feat: latest RHEL7.x and 8.x support
• fix: add jq to setup-system
• fix: conda version comparison fixed for fresh install, rm miniconda.sh temp file
• fix: remove docker-adduser command, replaced with setup-user
• docs: Document how to remove environments and containers
• docs: Add section on validating stack and notebook testing
• fix: yum failures not treated as fatal
• feat: Support for Ubuntu 20.04, update to 450 drivers
• feat: Update default stacks to RAPIDS 0.14.1 and re-pin stacks
• docs: document upgrade process and effects
• docs: Document creation of custom stacks and debug process
• docs: Document how to report issues and feature requests
• docs: Add troubleshooting section with common errors
• feat: Add commands to remove the container and conda environment
• feat: don't require setup-user as a seperate step for initial user
• feat: Additional diagnostics
• feat: Have version command check GitHub for new releases
• feat: If older Conda exists, ask user to upgrade instead of new install
• feat: Update default stacks to RAPIDS 0.14, version to 2.4.0-dev

v2.3.1

feat: display working dir and script when telling user to reboot or logout
fix: detect missing nvidia-smi

v2.3.0

Since 2.2.2:
feat: 2.3 environment files and pinned yamls
fix: remove ipyvolume from labextensions
docs: Update README for 2.3
feat: use channels from .env file in pinned yaml
docs: add list command to help, and display default
feat: on Ubuntu detect .run file was used and guide user to solution
fix: change how notebook repo is cloned and submodules handled.
feat: update versions to 2.3.0 and RAPIDS 0.13.
feat: add list command
fix: display default stack in help message
feat: log command used to pin, output to -pinned.yaml
docs: add CONTRIBUTING.md