Dataset Viewer issue: DatasetWithScriptNotSupportedError

#1
by Beijuka - opened

The dataset viewer is not working.

Error details:

Error code:   DatasetWithScriptNotSupportedError

No (supported) data files found in DigitalUmuganda/image_text_voice_dataset


πŸ‘‹ Before opening the discussion, have you considered removing the loading script and relying on [automated data support](
Screenshot from 2024-06-14 13-55-21.png
)?

You can use convert_to_parquet from the datasets library.


cc @albertvillanova @lhoestq @severo .

Digital Umuganda org

Hello, make sure that you are logged in huggingface (huggingface-cli login), and then use data = load_dataset("DigitalUmuganda/AfriVoice", "sn") to load the dataset.

rutsam changed discussion status to closed

Hello Kleber,
Thank you for the clarification from Digital Umuganda org. I have been trying to do some EDA on the shona dataset but cannot access the audio files. It seems the audio files were not added.
If possible maybe you can clarify on the structure of the data.
regards,
Bruno Beijuka

rutsam changed discussion status to open

Hello Kleber,
Thank you for the clarification from Digital Umuganda org. I have been trying to do some EDA on the shona dataset but cannot access the audio files. It seems the audio files were not added.
If possible maybe you can clarify on the structure of the data.
regards,
Bruno Beijuka

Digital Umuganda org

Hello Beijuka, the dataset is not loaded in the memory due to its size. So to access it after loading it with load_dataset; you can look for its location on the disk (for example: 'audio_path': '/root/.cache/huggingface/datasets/downloads/extracted/9347eb035e3ae38aaf793efa152ba1c93a4336471afce2bbd00ac8c0f67e9066/small_data/audio/I7L1YJVKIRL4.wav').

Sign up or log in to comment