EuroEval 15.3.0__py3-none-any.whl → 15.3.1__py3-none-any.whl

euroeval/benchmarker.py

@@ -18,7 +18,7 @@ from .data_loading import load_data
 from .data_models import BenchmarkConfigParams, BenchmarkResult
 from .dataset_configs import get_all_dataset_configs
 from .enums import Device, ModelType
-from .exceptions import InvalidBenchmark, InvalidModel
+from .exceptions import HuggingFaceHubDown, InvalidBenchmark, InvalidModel
 from .finetuning import finetune
 from .generation import generate
 from .model_config import get_model_config
@@ -769,23 +769,21 @@ class Benchmarker:
                 logger.debug(f"Results:\n{results}")
                 return record
 
+            except HuggingFaceHubDown:
+                wait_time = 30
+                logger.debug(
+                    f"The Hugging Face Hub seems to be down. Retrying in {wait_time} "
+                    "seconds."
+                )
+                sleep(wait_time)
+                continue
+
             except (InvalidBenchmark, InvalidModel) as e:
                 # If the model ID is not valid then raise an error
                 model_err_msg = "does not exist on the Hugging Face Hub"
                 if benchmark_config.raise_errors and model_err_msg in str(e):
                     raise e
 
-                # Otherwise, if the error is due to Hugging Face Hub being down, then
-                # wait a bit and try again
-                elif "The Hugging Face Hub seems to be down." in str(e):
-                    wait_time = 30
-                    logger.debug(
-                        "The Hugging Face Hub seems to be down. Retrying in "
-                        f"{wait_time} seconds."
-                    )
-                    sleep(wait_time)
-                    continue
-
                 # Otherwise, if the error is due to the MPS fallback not being enabled,
                 # then raise an error asking the user to enable it
                 elif "PYTORCH_ENABLE_MPS_FALLBACK" in str(e):

euroeval/data_loading.py

@@ -10,7 +10,7 @@ from huggingface_hub.errors import HfHubHTTPError
 from numpy.random import Generator
 
 from .data_models import BenchmarkConfig, DatasetConfig
-from .exceptions import InvalidBenchmark
+from .exceptions import HuggingFaceHubDown, InvalidBenchmark
 from .utils import unscramble
 
 logger = logging.getLogger("euroeval")
@@ -31,6 +31,12 @@ def load_data(
 
     Returns:
         A list of bootstrapped datasets, one for each iteration.
+
+    Raises:
+        InvalidBenchmark:
+            If the dataset cannot be loaded.
+        HuggingFaceHubDown:
+            If the Hugging Face Hub is down.
     """
     num_attempts = 5
     for _ in range(num_attempts):
@@ -41,14 +47,14 @@ def load_data(
                 token=unscramble("HjccJFhIozVymqXDVqTUTXKvYhZMTbfIjMxG_"),
             )
             break
-        except (FileNotFoundError, DatasetsError):
+        except (FileNotFoundError, DatasetsError, ConnectionError):
             logger.warning(
                 f"Failed to load dataset {dataset_config.huggingface_id!r}. Retrying..."
             )
             time.sleep(1)
             continue
         except HfHubHTTPError:
-            raise InvalidBenchmark("The Hugging Face Hub seems to be down.")
+            raise HuggingFaceHubDown()
     else:
         raise InvalidBenchmark(
             f"Failed to load dataset {dataset_config.huggingface_id!r} after "

{euroeval-15.3.0.dist-info → euroeval-15.3.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: EuroEval
-Version: 15.3.0
+Version: 15.3.1
 Summary: The robust European language model benchmark.
 Project-URL: Repository, https://github.com/EuroEval/EuroEval
 Project-URL: Issues, https://github.com/EuroEval/EuroEval/issues

{euroeval-15.3.0.dist-info → euroeval-15.3.1.dist-info}/RECORD

@@ -1,10 +1,10 @@
 euroeval/__init__.py,sha256=3od9_ucHlILSbe4WCR8k5PbeorvmUr-VjOKXJ01I0fA,2165
 euroeval/benchmark_config_factory.py,sha256=pi4Lu--ySKZRd9ItG6VKS6BPLis64vL-7UE99VSXq5Y,12534
-euroeval/benchmarker.py,sha256=EjORG5haUio9LgfGH7ruWEFutvJN0QGasoknFH_yGHs,46705
+euroeval/benchmarker.py,sha256=__DdnOvI9CNpgqPT1hsTl0GZFTyQ6KRfiQowCuh36sc,46534
 euroeval/callbacks.py,sha256=bThUUxOgkMuESUQ5rrFRoSumKV8vNw53CslIZTpkt54,2438
 euroeval/cli.py,sha256=EMB6g6kRvxIqlfYLSoMzwLAtEd-fqXipo4A_HTkhjkA,8575
 euroeval/constants.py,sha256=qFrm3cRT6UlnTXfHUmxqZsr0SBsGskjV1qrUlnAW-aw,1473
-euroeval/data_loading.py,sha256=IHd1H4OCAtOyiro7YnJsGbbT7PTwiMUB02gh1g6Nlhg,3116
+euroeval/data_loading.py,sha256=RoatBJMpGurP_y5O3KrEvly8Z_yYEapQnnMZ_tWWrlc,3272
 euroeval/data_models.py,sha256=4ZY9x2pINlRywTzYxxtrYG7qXMNdod5I9XBOlTJYT8E,14495
 euroeval/dataset_configs.py,sha256=Cj3McxA0JTC7RKzXofzpJfmIhoXAfF756f_1SZUaPlw,84391
 euroeval/enums.py,sha256=L9LcNeruuhHvze9vKRogXY9vonRzoBqDzWSP6hxKQ7A,3195
@@ -33,8 +33,8 @@ euroeval/task_utils/question_answering.py,sha256=G01s11JcQ7UxeBcKaCO3k0DL4zkVmEb
 euroeval/task_utils/sequence_classification.py,sha256=FrkvFzxFSnZoXThgpQqvJCIy3_YemyqZFQ1L-YdMMiw,8527
 euroeval/task_utils/text_to_text.py,sha256=DdLruAO4D9Iv5aAXx40la3X3pKbKLUn0-ViBJkMKsTI,5698
 euroeval/task_utils/token_classification.py,sha256=yT1YvZzmqNaVSRZ67BvyURhlkgTm3ltWPft4HxodZAE,17983
-euroeval-15.3.0.dist-info/METADATA,sha256=Mlz6DcLg2H3aWoCXngQZNdFMrJmUFpAdD0FD0wsBKHw,10263
-euroeval-15.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-euroeval-15.3.0.dist-info/entry_points.txt,sha256=tKQRxN0HX2mGtbZbZQdCRFUDZIecA_z4mZduueor3Ug,135
-euroeval-15.3.0.dist-info/licenses/LICENSE,sha256=oZp5fpOSQ7w-vFui8QNwrBIosrO7cnpArItdbvn52Ao,1082
-euroeval-15.3.0.dist-info/RECORD,,
+euroeval-15.3.1.dist-info/METADATA,sha256=elF7s_zt2tj9Hl1EMMDfNoMtskYK5Xh9i-N36vvzfQs,10263
+euroeval-15.3.1.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+euroeval-15.3.1.dist-info/entry_points.txt,sha256=tKQRxN0HX2mGtbZbZQdCRFUDZIecA_z4mZduueor3Ug,135
+euroeval-15.3.1.dist-info/licenses/LICENSE,sha256=oZp5fpOSQ7w-vFui8QNwrBIosrO7cnpArItdbvn52Ao,1082
+euroeval-15.3.1.dist-info/RECORD,,