v2026.1.1

[djgagne]

Summary

• New users can now train and run a model based entirely on the config/example-v2026.1.0.yml config file on Casper or Derecho without having to make modifications. For users on other systems, data paths will need to change but other components will still work.
• This is the last release focused on the CREDIT Generation 1 API. v2026.2.0 will begin implementing the Generation 2 API, which will use a more modular set of multi-source Datasets, pre-blocks, and post-blocks to handle data transformations and diagnostic calculations. The Generation 1 API and associated dataset classes and models will still be maintained in the new update but future feature development will focus on the Gen 2 API. Conversion scripts for config files to go from Gen 1 to Gen 2 will also be provided.
• New Gen 2 Datasets and Preblocks are included in this release but are not fully connected with the model pipeline yet. We will have a full Gen 2 model pipeline in the next release.
• We highly recommend building a new conda environment before running this CREDIT release. The number of dependencies has been reduced, and the code has been tested against numpy 2.4 and the most recent versions of pytorch. See the Getting Started instructions in the docs on how to create a new environment and link to a CUDA-enabled version of PyTorch that can run on Casper or other NVIDIA GPU machines. If you want to use the p-type functionality or any of the optional dependencies, run pip install -e ".[ptype]" or pip install miles-credit[ptype]. We also have a [develop] set of optional dependencies if you are making code contributions.
• The Derecho environment creation script has also been updated to support the newer versions of pytorch compiled with the system MKL and MPI.
• We have completed a bunch of updates under the hood to improve testing and code contributing with many more planned in the next few releases.

What's Changed

• Streamlines Ruff configuration, adds pre-commit.ci, and applies linting/formatting fixes by @kafitzgerald in Streamlines Ruff configuration, adds pre-commit.ci, and applies linting/formatting fixes #269
• Fix MPI conda channel issue in create_derecho_env.sh by @djgagne in Fix MPI conda channel issue in create_derecho_env.sh #279
• two fixes -- mpi + torch version for derecho by @negin513 in two fixes -- mpi + torch version for derecho #281
• Update era5.py by @charlie-becker in Update era5.py #280
• Update getting-started.md by @sgpearse in Update getting-started.md #282
• Bump the actions group with 2 updates by @dependabot[bot] in Bump the actions group with 2 updates #284
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in [pre-commit.ci] pre-commit autoupdate #285
• Preblocks by @charlie-becker in Preblocks #283
• Dataset refactor by @charlie-becker in Dataset refactor #291
• Bump codecov/codecov-action from 5.5.2 to 5.5.3 in the actions group by @dependabot[bot] in Bump codecov/codecov-action from 5.5.2 to 5.5.3 in the actions group #293
• Fix to example config by @djgagne in Fix to example config #294
• CI check to run inference on each Pull Request by @sgpearse in CI check to run inference on each Pull Request #286
• adding a new energy post-block that doesn't use net notation by @WillyChap in adding a new energy post-block that doesn't use net notation #301
• Bump the actions group with 2 updates by @dependabot[bot] in Bump the actions group with 2 updates #311
• Coupled camulator by @WillyChap in Coupled camulator #287
• CI / Actions updates and bump minimum Python version to 3.11 by @kafitzgerald in CI / Actions updates and bump minimum Python version to 3.11 #313
• Updated example config and docs by @djgagne in Updated example config and docs #300
• Bump actions/download-artifact from 8.0.0 to 8.0.1 in the actions group across 1 directory by @dependabot[bot] in Bump actions/download-artifact from 8.0.0 to 8.0.1 in the actions group across 1 directory #321
• Domain parallelism for V2: spatial sharding across GPUs for high-resolution training by @negin513 in Domain parallelism for V2: spatial sharding across GPUs for high-resolution training #320
• Add a PR template by @kafitzgerald in Add a PR template #325
• [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in [pre-commit.ci] pre-commit autoupdate #326
• Preblocks by @charlie-becker in Preblocks #297
• Configuration File Docs Text Update by @danineamati in Configuration File Docs Text Update #327
• Bump the actions group across 1 directory with 2 updates by @dependabot[bot] in Bump the actions group across 1 directory with 2 updates #336
• gfs_init and rollout_realtime now work with the example config by @djgagne in gfs_init and rollout_realtime now work with the example config #334
• Deleting old notebooks and large files in repo. by @djgagne in Deleting old notebooks and large files in repo. #338

New Contributors

• @pre-commit-ci[bot] made their first contribution in [pre-commit.ci] pre-commit autoupdate #285

Full Changelog: v2026.1.0...v2026.1.1

---

This discussion was created from the release v2026.1.1.