README.md

metadata

license: openrail
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: label
      dtype: int64
    - name: is_unknown
      dtype: bool
    - name: speaker_id
      dtype: string
    - name: utterance_id
      dtype: int8
    - name: logits
      sequence: float32
    - name: Probability
      dtype: float64
    - name: Predicted Label
      dtype: string
    - name: Annotated Labels
      dtype: string
    - name: embedding
      sequence: float32
    - name: embedding_reduced
      sequence: float32
  splits:
    - name: train
      num_bytes: 2137341907.375
      num_examples: 60973
  download_size: 2094280286
  dataset_size: 2137341907.375

This dataset is an extended version of the MIT/ast-finetuned-speech-commands-v2 dataset. It provides predicted labels, their annotations and embeddings, trained with Huggingface's AutoModel and AutoFeatureExtractor. If you would like to have a closer look at the dataset and model's performance, you can use Spotlight by Renumics to find complex sub-relationships between classes.