Hi, recently I am having a problem with dataset viewer. I cannot get the preview of audio files in my dataset. I am getting this error :
ERROR: type should be audio, got [ { "src": "https://datasets-server.huggingface.co/assets/j-krzywdziak/test/--/Test Dataset/test/2/audio/audio.mp3", "type": "audio/mpeg" }, { "src": "https://datasets-server.huggingface.co/assets/j-krzywdziak/test/--/Test Dataset/test/2/audio/audio.wav", "type": "audio/wav" } ]
Audio files are zipped and are in wav format. Anyone have any idea?
Hi ! Which dataset is it ? The audio type must contain either
- a list of strings pointing to the WAV files
- a list of dictionaries with “bytes” and “path” field - containing the file bytes and the file name or path
Hi! it is j-krzywdziak/test · Datasets at Hugging Face , In test.py · j-krzywdziak/test at main in _generate_examples i added path and bytes but something gone wrong along the way.
@lhoestq Do you have any idea why do I have this error? I would appreciate any help because i spent almost 3 days trying to solve this ;p
Your code looks all good to me, this sounds like a bug on our side. I’m forcing a refresh of your dataset preview. If it persists, please open an issue at GitHub - huggingface/datasets-server: Lightweight web API for visualizing and exploring all types of datasets - computer vision, speech, text, and tabular - stored on the Hugging Face Hub
1 Like
created Dataset Viewer issue for j-krzywdziak/test · Issue #966 · huggingface/datasets-server · GitHub
@severo fyi - it’s a duplicate of Problem with Dataset Preview with audio files and the error is caused by a space symbol in the config name, you already opened an issue for that 
2 Likes