Dataset Viewer issue: ConfigNamesError

#1
by evie-8 - opened

The dataset viewer is not working.

Error details:

Error code:   ConfigNamesError
Exception:    RuntimeError
Message:      Dataset scripts are no longer supported, but found Afrivoice_Swahili.py
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/dataset/config_names.py", line 66, in compute_config_names_response
                  config_names = get_dataset_config_names(
                                 ^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 161, in get_dataset_config_names
                  dataset_module = dataset_module_factory(
                                   ^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/load.py", line 1029, in dataset_module_factory
                  raise e1 from None
                File "/usr/local/lib/python3.12/site-packages/datasets/load.py", line 989, in dataset_module_factory
                  raise RuntimeError(f"Dataset scripts are no longer supported, but found {filename}")
              RuntimeError: Dataset scripts are no longer supported, but found Afrivoice_Swahili.py

cc @albertvillanova @lhoestq @severo .

Digital Umuganda org

For the dataset scripts to work, you need to downgrade the datasets library to datasets==3.6.0

The latest versions don't allow dataset scripts. Or you can download using huggingface_hub, but you will have to unzip all the files and create a data loader for the model training.

Thank you for the clarification.

However, it would be highly beneficial if the dataset were provided in a format that aligns with the current Hugging Face datasets standards. Requiring users to downgrade from newer versions (e.g., 4.x) to 3.6.0 for a single dataset introduces avoidable friction and may impact compatibility with other dependencies in existing workflows.

Adopting a natively supported format—such as distributing the data via huggingface_hub with a standard structure or loader—would make the dataset significantly easier to use, integrate, and maintain within modern pipelines.

Addressing this would improve accessibility and overall usability for a broader range of users.

Sign up or log in to comment