metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: text_instruction
dtype: string
- name: data_index
dtype: string
- name: relevance_score
dtype: int32
- name: quality_score
dtype: int32
- name: relevance_reasoning
dtype: string
- name: quality_reasoning
dtype: string
- name: segments_json
dtype: string
- name: origin_model
dtype: string
- name: origin_config
dtype: string
splits:
- name: train
num_bytes: 22618291
num_examples: 150
download_size: 21492336
dataset_size: 22618291
configs:
- config_name: default
data_files:
- split: train
path: data/train-*