Datasets:

Languages:
English
ArXiv:
License:

Dataset Viewer issue: DatasetWithScriptNotSupportedError

#4
by evie-8 - opened

The dataset viewer is not working.

Error details:

Error code:   DatasetWithScriptNotSupportedError

👋 Before opening the discussion, have you considered removing the loading script and relying on automated data support?

You can use convert_to_parquet from the datasets library.


cc @albertvillanova @lhoestq @severo .

Hi ! This dataset is in an unsupported format, but you can convert it to Parquet yourself by running the datasets-cli convert_to_parquet command with uv:

uv run --with "datasets @ git+https://github.com/huggingface/datasets@large_convert_to_parquet" \
  datasets-cli convert_to_parquet intronhealth/afrispeech-200 --trust_remote_code

You can also run it on Hugging Face using HF Jobs:

hf jobs uv run --flavor cpu-performance --timeout 4d -s HF_TOKEN \
  --with "datasets @ git+https://github.com/huggingface/datasets@large_convert_to_parquet" \
  datasets-cli convert_to_parquet intronhealth/afrispeech-200 --trust_remote_code

This will open a PR with the Parquet version. Once merged, the Dataset Viewer will be available and the dataset will be supported in most data libraries like pandas, dask, etc.

evie-8 changed discussion status to closed

Sign up or log in to comment