ScandEval 16.10.0__tar.gz → 16.10.1__tar.gz

{scandeval-16.10.0 → scandeval-16.10.1}/.pre-commit-config.yaml

@@ -34,7 +34,7 @@ repos:
     hooks:
     -   id: nbstripout
 -   repo: https://github.com/facebook/pyrefly-pre-commit
-    rev: 0.46.2
+    rev: 0.46.3
     hooks:
     -   id: pyrefly-check
         name: Pyrefly (type checking)

{scandeval-16.10.0 → scandeval-16.10.1}/CHANGELOG.md

@@ -7,6 +7,19 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ## [Unreleased]
 
+## [v16.10.1] - 2026-01-02
+
+### Changed
+
+- Now set the `duidelijke-taal` dataset as official for Dutch simplification, as it is
+  now included on the leaderboards now. It will thus automatically be run when
+  evaluating any generative model in Dutch.
+
+### Fixed
+
+- Nynorsk datasets were incorrectly not included when evaluating models in Norwegian.
+  This has been fixed now.
+
 ## [v16.10.0] - 2025-12-30
 
 ### Added

{scandeval-16.10.0 → scandeval-16.10.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ScandEval
-Version: 16.10.0
+Version: 16.10.1
 Summary: The robust European language model benchmark.
 Project-URL: Repository, https://github.com/EuroEval/EuroEval
 Project-URL: Issues, https://github.com/EuroEval/EuroEval/issues

{scandeval-16.10.0 → scandeval-16.10.1}/docs/datasets/dutch.md

@@ -1031,7 +1031,7 @@ euroeval --model <model-id> --dataset wiki-lingua-nl
 
 ## Simplification
 
-### Unofficial: Duidelijke Taal
+### Duidelijke Taal
 
 The [Duidelijke Taal dataset](http://hdl.handle.net/10032/tm-a2-y8) was created by
 Instituut voor de Nederlandse Taal and published in _Human Evaluation of Automated Text

{scandeval-16.10.0 → scandeval-16.10.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "ScandEval"
-version = "16.10.0"
+version = "16.10.1"
 description = "The robust European language model benchmark."
 readme = "README.md"
 authors = [

{scandeval-16.10.0 → scandeval-16.10.1}/src/scandeval/benchmark_modules/vllm.py

@@ -1454,10 +1454,11 @@ def select_backend_and_parallelism() -> tuple[str, int, int]:
         try:
             ray.init(address="auto", ignore_reinit_error=True)
         except Exception as e:
-            log_once(
-                f"Ray initialisation failed with a {type(e)} exception: {e}",
-                level=logging.DEBUG,
-            )
+            if "could not find any running ray instance" not in str(e).lower():
+                log_once(
+                    f"Ray initialisation failed with a {type(e)} exception: {e}",
+                    level=logging.DEBUG,
+                )
 
     is_ray = ray.is_initialized()
     local_gpu_count = torch.cuda.device_count()

{scandeval-16.10.0 → scandeval-16.10.1}/src/scandeval/dataset_configs/dutch.py

@@ -74,6 +74,14 @@ HELLASWAG_NL_CONFIG = DatasetConfig(
     languages=[DUTCH],
 )
 
+DUIDELIJKE_TAAL_NL_CONFIG = DatasetConfig(
+    name="duidelijke-taal",
+    pretty_name="Duidelijke Taal",
+    source="EuroEval/duidelijke-taal",
+    task=SIMPL,
+    languages=[DUTCH],
+)
+
 VALEU_NL_CONFIG = DatasetConfig(
     name="valeu-nl",
     pretty_name="VaLEU-nl",
@@ -161,12 +169,3 @@ WINOGRANDE_NL_CONFIG = DatasetConfig(
     _labels=["a", "b"],
     unofficial=True,
 )
-
-DUIDELIJKE_TAAL_NL_CONFIG = DatasetConfig(
-    name="duidelijke-taal",
-    pretty_name="Duidelijke Taal",
-    source="EuroEval/duidelijke-taal",
-    task=SIMPL,
-    languages=[DUTCH],
-    unofficial=True,
-)

{scandeval-16.10.0 → scandeval-16.10.1}/src/scandeval/dataset_configs/norwegian.py

@@ -27,7 +27,7 @@ SCALA_NN_CONFIG = DatasetConfig(
     pretty_name="ScaLA-nn",
     source="EuroEval/scala-nn",
     task=LA,
-    languages=[NORWEGIAN_NYNORSK],
+    languages=[NORWEGIAN_NYNORSK, NORWEGIAN],
 )
 
 NORNE_NB_CONFIG = DatasetConfig(
@@ -43,7 +43,7 @@ NORNE_NN_CONFIG = DatasetConfig(
     pretty_name="NorNE-nn",
     source="EuroEval/norne-nn-mini",
     task=NER,
-    languages=[NORWEGIAN_NYNORSK],
+    languages=[NORWEGIAN_NYNORSK, NORWEGIAN],
 )
 
 NORQUAD_CONFIG = DatasetConfig(
@@ -197,7 +197,7 @@ MULTI_WIKI_QA_NN_CONFIG = DatasetConfig(
     pretty_name="MultiWikiQA-nn",
     source="EuroEval/multi-wiki-qa-nn-mini",
     task=RC,
-    languages=[NORWEGIAN_NYNORSK],
+    languages=[NORWEGIAN_NYNORSK, NORWEGIAN],
     unofficial=True,
 )
 

{scandeval-16.10.0 → scandeval-16.10.1}/uv.lock

@@ -980,7 +980,7 @@ wheels = [
 
 [[package]]
 name = "euroeval"
-version = "16.10.0"
+version = "16.10.1"
 source = { editable = "." }
 dependencies = [
     { name = "accelerate" },