split
string
total_samples
int64
failed_samples
int64
samples
list
train
21,557
43
[ { "duration": 4.7934375, "features_file": "sample_0_features.npy", "index": 0, "labels_file": "sample_0_labels.npy", "num_labels": 21, "source": "Mira", "speaker_id": 1, "text": "If what I say comes true, it is for you what to do, said Sham." }, { "duration": 4.65210416666666...