Clotho dataset in WebDataset Format

This dataset is the Clotho dataset, formatted in the WebDataset format. WebDataset files are essentially tar archives, where each example in the dataset is represented by a pair of files: a WAV audio file and a corresponding JSON metadata file. The JSON file contains the class label and other relevant information for that particular audio sample.

$ tar tvf clotho_development_0000000.tar|head                                                                    
-r--r--r-- bigdata/bigdata 386 2025-01-14 08:59 Toilet Flush with Cistern.json
-r--r--r-- bigdata/bigdata 2328620 2025-01-14 08:59 Toilet Flush with Cistern.wav
-r--r--r-- bigdata/bigdata     360 2025-01-14 08:59 Mouth Breating.json
-r--r--r-- bigdata/bigdata 1897516 2025-01-14 08:59 Mouth Breating.wav
-r--r--r-- bigdata/bigdata     500 2025-01-14 08:59 STE-040 italian police siren.json
-r--r--r-- bigdata/bigdata 1534562 2025-01-14 08:59 STE-040 italian police siren.wav
-r--r--r-- bigdata/bigdata     378 2025-01-14 08:59 Water Sound.json
-r--r--r-- bigdata/bigdata 2070316 2025-01-14 08:59 Water Sound.wav
-r--r--r-- bigdata/bigdata     373 2025-01-14 08:59 RoostingBirds.json
-r--r--r-- bigdata/bigdata 1607966 2025-01-14 08:59 RoostingBirds.wav

$ tar -xOf clotho_development_0000000.tar RoostingBirds.json
{
      "caption_1": "A large amount of birds are chirping while cars pass by in the background.", 
      "caption_2": "Many birds are chirping and cars are driving by quickly.", 
      "caption_3": "Many birds are chirping and cars are driving by.", 
      "caption_4": "Outside, many types of different birds are tweeting.", 
      "caption_5": "bird chirping loudly with traffic sounds in the background"
}