metadata
dataset_info:
features:
- name: sample_id
dtype: string
- name: dataset_id
dtype: string
- name: subset_id
dtype: string
- name: reference
dtype: string
- name: hypothesis
dtype: string
- name: system_id
dtype: string
- name: wer
dtype: float64
- name: cer
dtype: float64
- name: error_type
dtype: string
- name: severity
dtype: int64
- name: audio_quality_ok
dtype: bool
- name: reference_correct
dtype: bool
- name: notes
dtype: string
- name: annotator_id
dtype: string
- name: llm_error_categories
sequence: string
- name: llm_severity
dtype: string
- name: llm_confidence
dtype: float64
- name: llm_model_id
dtype: string
- name: split
dtype: string
splits:
- name: train
num_bytes: 4674
num_examples: 11
download_size: 11429
dataset_size: 4674
configs:
- config_name: default
data_files:
- split: train
path: data/train-*