atmatechdev/dataset_30k_without_norm_pseudo_labelled_medium

---
dataset_info:
config_name: id
features:
- name: audio
dtype:
audio:
sampling_rate: 16000
- name: sentence
dtype: string
- name: path
dtype: string
- name: condition_on_prev
sequence: int64
- name: whisper_transcript
dtype: string
splits:
- name: train
num_bytes: 4304850996.048
num_examples: 23952
- name: test
num_bytes: 1079939484.512
num_examples: 5988
download_size: 5296637806
dataset_size: 5384790480.56
configs:
- config_name: id
data_files:
- split: train
path: id/train-*
- split: test
path: id/test-*
---

原始信息汇总

数据集概述

数据集配置名称

• config_name: id

数据集特征

• audio:
  • dtype:
    • sampling_rate: 16000
• sentence:
  • dtype: string
• path:
  • dtype: string
• condition_on_prev:
  • sequence: int64
• whisper_transcript:
  • dtype: string

数据集分割

• train:
  • num_examples: 23952
  • num_bytes: 4304850996.048
• test:
  • num_examples: 5988
  • num_bytes: 1079939484.512

数据集大小

• download_size: 5296637806
• dataset_size: 5384790480.56

数据文件路径

• train: id/train-*
• test: id/test-*