SALMon_pgslm_v2 / README.md
dlion168's picture
Upload dataset
297a7d7 verified
metadata
dataset_info:
  - config_name: bg_alignment
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        dtype: 'null'
      - name: negative_continuation_tokenwise_loss
        dtype: 'null'
      - name: postive_continuation_raw_units
        dtype: 'null'
      - name: negative_continuation_raw_units
        dtype: 'null'
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 99501618
        num_examples: 200
    download_size: 87225435
    dataset_size: 99501618
  - config_name: bg_all_consistency
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        list: float64
      - name: negative_continuation_tokenwise_loss
        list: float64
      - name: postive_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: negative_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 207692462
        num_examples: 200
    download_size: 207264432
    dataset_size: 207692462
  - config_name: bg_domain_consistency
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        list: float64
      - name: negative_continuation_tokenwise_loss
        list: float64
      - name: postive_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: negative_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 210848264
        num_examples: 200
    download_size: 210423859
    dataset_size: 210848264
  - config_name: gender_consistency
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        list: float64
      - name: negative_continuation_tokenwise_loss
        list: float64
      - name: postive_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: negative_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 214533433
        num_examples: 200
    download_size: 208860886
    dataset_size: 214533433
  - config_name: rir_consistency
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        list: float64
      - name: negative_continuation_tokenwise_loss
        list: float64
      - name: postive_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: negative_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 208070335
        num_examples: 200
    download_size: 207297739
    dataset_size: 208070335
  - config_name: sentiment_alignment
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        dtype: 'null'
      - name: negative_continuation_tokenwise_loss
        dtype: 'null'
      - name: postive_continuation_raw_units
        dtype: 'null'
      - name: negative_continuation_raw_units
        dtype: 'null'
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 56840084
        num_examples: 200
    download_size: 46778551
    dataset_size: 56840084
  - config_name: sentiment_consistency
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        list: float64
      - name: negative_continuation_tokenwise_loss
        list: float64
      - name: postive_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: negative_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 211263288
        num_examples: 200
    download_size: 210115475
    dataset_size: 211263288
  - config_name: speaker_consistency
    features:
      - name: task
        dtype: string
      - name: ind
        dtype: int64
      - name: postive_sample_tokenwise_loss
        list: float32
      - name: negative_sample_tokenwise_loss
        list: float32
      - name: prompt_sample_tokenwise_loss
        list: float32
      - name: model_generated_continuation
        dtype:
          audio:
            sampling_rate: 16000
      - name: positive_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: negative_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: postive_sample_raw_units
        list: string
      - name: negative_sample_raw_units
        list: string
      - name: prompt_audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: code_frame_rate
        dtype: int64
      - name: code_depth
        dtype: int64
      - name: model_sampling_rate
        dtype: int64
      - name: ppl_sanity
        dtype: int64
      - name: postive_continuation_tokenwise_loss
        list: float64
      - name: negative_continuation_tokenwise_loss
        list: float64
      - name: postive_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: negative_continuation_raw_units
        list:
          - name: duration
            dtype: string
          - name: hubert
            dtype: string
          - name: pitch
            dtype: string
      - name: continuation_audio_positive
        dtype:
          audio:
            sampling_rate: 16000
      - name: continuation_audio_negative
        dtype:
          audio:
            sampling_rate: 16000
    splits:
      - name: train
        num_bytes: 215625692
        num_examples: 200
    download_size: 209872318
    dataset_size: 215625692
configs:
  - config_name: bg_alignment
    data_files:
      - split: train
        path: bg_alignment/train-*
  - config_name: bg_all_consistency
    data_files:
      - split: train
        path: bg_all_consistency/train-*
  - config_name: bg_domain_consistency
    data_files:
      - split: train
        path: bg_domain_consistency/train-*
  - config_name: gender_consistency
    data_files:
      - split: train
        path: gender_consistency/train-*
  - config_name: rir_consistency
    data_files:
      - split: train
        path: rir_consistency/train-*
  - config_name: sentiment_alignment
    data_files:
      - split: train
        path: sentiment_alignment/train-*
  - config_name: sentiment_consistency
    data_files:
      - split: train
        path: sentiment_consistency/train-*
  - config_name: speaker_consistency
    data_files:
      - split: train
        path: speaker_consistency/train-*