From 100a6ba7998b7beadd254f38655766ebc723eb92 Mon Sep 17 00:00:00 2001 From: ECMWF <150700357+DeployDuck@users.noreply.github.com> Date: Fri, 22 May 2026 13:17:44 +0100 Subject: [PATCH 1/2] chore: Release main --- .release-please-manifest.json | 4 ++-- models/CHANGELOG.md | 18 ++++++++++++++++++ training/CHANGELOG.md | 22 ++++++++++++++++++++++ 3 files changed, 42 insertions(+), 2 deletions(-) diff --git a/.release-please-manifest.json b/.release-please-manifest.json index 0e2523c74b..50360dd487 100644 --- a/.release-please-manifest.json +++ b/.release-please-manifest.json @@ -1,5 +1,5 @@ { - "training": "0.13.0", + "training": "0.14.0", "graphs": "0.9.4", - "models": "0.15.1" + "models": "0.16.0" } diff --git a/models/CHANGELOG.md b/models/CHANGELOG.md index fe50bb9683..76f6e12cfc 100644 --- a/models/CHANGELOG.md +++ b/models/CHANGELOG.md @@ -8,6 +8,24 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 Please add your functional changes to the appropriate section in the PR. Keep it human-readable, your future self will thank you! +## [0.16.0](https://github.com/ecmwf/anemoi-core/compare/models-0.15.1...models-0.16.0) (2026-05-22) + + +### ⚠ BREAKING CHANGES + +* introduce mlp implementation for gated FFN (swiglu, ...) ([#943](https://github.com/ecmwf/anemoi-core/issues/943)) + +### Features + +* Introduce mlp implementation for gated FFN (swiglu, ...) ([#943](https://github.com/ecmwf/anemoi-core/issues/943)) ([2ed3417](https://github.com/ecmwf/anemoi-core/commit/2ed3417906423cd8be7ee432138b94606deaacb9)) + + +### Bug Fixes + +* Changed FFT2D low-pass filter from enabled to disabled by default ([#1075](https://github.com/ecmwf/anemoi-core/issues/1075)) ([336bacf](https://github.com/ecmwf/anemoi-core/commit/336bacf65244039f944076711b1e98466c24249f)) +* **models:** Gather_in_backward=False ([#1138](https://github.com/ecmwf/anemoi-core/issues/1138)) ([b732f6a](https://github.com/ecmwf/anemoi-core/commit/b732f6af39841e36b8a9bc9c9e29e2caf1cce832)) +* **models:** Hierarchical model with multiple datasets ([#1066](https://github.com/ecmwf/anemoi-core/issues/1066)) ([987aa09](https://github.com/ecmwf/anemoi-core/commit/987aa09dba164f3f447e8391a2ea7a3c074fed4e)) + ## [0.15.1](https://github.com/ecmwf/anemoi-core/compare/models-0.15.0...models-0.15.1) (2026-05-15) diff --git a/training/CHANGELOG.md b/training/CHANGELOG.md index 234b28305e..a88752b0a0 100644 --- a/training/CHANGELOG.md +++ b/training/CHANGELOG.md @@ -8,6 +8,28 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 Please add your functional changes to the appropriate section in the PR. Keep it human-readable, your future self will thank you! +## [0.14.0](https://github.com/ecmwf/anemoi-core/compare/training-0.13.0...training-0.14.0) (2026-05-22) + + +### ⚠ BREAKING CHANGES + +* aggregate losses for temporal downscaler ([#1069](https://github.com/ecmwf/anemoi-core/issues/1069)) +* introduce mlp implementation for gated FFN (swiglu, ...) ([#943](https://github.com/ecmwf/anemoi-core/issues/943)) + +### Features + +* Abstract asynchronous plotting executor ([#1129](https://github.com/ecmwf/anemoi-core/issues/1129)) ([f0743b7](https://github.com/ecmwf/anemoi-core/commit/f0743b78893f04bff18057b68ca81e17f3c777b4)) +* Aggregate losses for temporal downscaler ([#1069](https://github.com/ecmwf/anemoi-core/issues/1069)) ([390ce34](https://github.com/ecmwf/anemoi-core/commit/390ce34c1fbf89bf37cc8733cc500fe2bc7237ee)) +* Callbacks dependencies ([#1128](https://github.com/ecmwf/anemoi-core/issues/1128)) ([f791b75](https://github.com/ecmwf/anemoi-core/commit/f791b75d325e8fcdd9b6507510efa5cde0915d01)) +* Introduce mlp implementation for gated FFN (swiglu, ...) ([#943](https://github.com/ecmwf/anemoi-core/issues/943)) ([2ed3417](https://github.com/ecmwf/anemoi-core/commit/2ed3417906423cd8be7ee432138b94606deaacb9)) + + +### Bug Fixes + +* Asyncio loop thread management with Python 3.13 ([#1124](https://github.com/ecmwf/anemoi-core/issues/1124)) ([7c9c8aa](https://github.com/ecmwf/anemoi-core/commit/7c9c8aacd24fedb14f8fd3ca2c4ceb923f13fbc1)) +* **models:** Hierarchical model with multiple datasets ([#1066](https://github.com/ecmwf/anemoi-core/issues/1066)) ([987aa09](https://github.com/ecmwf/anemoi-core/commit/987aa09dba164f3f447e8391a2ea7a3c074fed4e)) +* **training:** Gather sharded batches if loss does not support sharding ([#1130](https://github.com/ecmwf/anemoi-core/issues/1130)) ([921358c](https://github.com/ecmwf/anemoi-core/commit/921358cc6bf85ee8137a3f5922bae7e03e473096)) + ## [0.13.0](https://github.com/ecmwf/anemoi-core/compare/training-0.12.1...training-0.13.0) (2026-05-14) From 13d31c1f4e77662e031de709b45dbd410f0f4400 Mon Sep 17 00:00:00 2001 From: DeployDuck <150700357+DeployDuck@users.noreply.github.com> Date: Fri, 22 May 2026 12:17:58 +0000 Subject: [PATCH 2/2] chore: Update checkpoint migrations placeholder to 0.16.0 --- .../migrations/scripts/1778688800_glu_mlp_implementation.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/models/src/anemoi/models/migrations/scripts/1778688800_glu_mlp_implementation.py b/models/src/anemoi/models/migrations/scripts/1778688800_glu_mlp_implementation.py index 8e048c9aaf..90e630f046 100644 --- a/models/src/anemoi/models/migrations/scripts/1778688800_glu_mlp_implementation.py +++ b/models/src/anemoi/models/migrations/scripts/1778688800_glu_mlp_implementation.py @@ -14,7 +14,7 @@ metadata = MigrationMetadata( versions={ "migration": "1.0.0", - "anemoi-models": "%NEXT_ANEMOI_MODELS_VERSION%", + "anemoi-models": "0.16.0", }, ) # <-- END DO NOT CHANGE