metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: topic
dtype: string
- name: initial_chat_prompt
dtype: string
- name: timestamp_ch_0
struct:
- name: word
list:
- name: start
dtype: float64
- name: end
dtype: float64
- name: start_offset
dtype: int64
- name: end_offset
dtype: int64
- name: word
dtype: string
- name: segment
list:
- name: start
dtype: float64
- name: end
dtype: float64
- name: start_offset
dtype: int64
- name: end_offset
dtype: int64
- name: segment
dtype: string
- name: timestamp_ch_1
struct:
- name: word
list:
- name: start
dtype: float64
- name: end
dtype: float64
- name: start_offset
dtype: int64
- name: end_offset
dtype: int64
- name: word
dtype: string
- name: segment
list:
- name: start
dtype: float64
- name: end
dtype: float64
- name: start_offset
dtype: int64
- name: end_offset
dtype: int64
- name: segment
dtype: string
- name: prompt
dtype: string
- name: gemini-2.5-flash_response
dtype: string
- name: tt_score
dtype: int64
- name: mt_score
dtype: int64
splits:
- name: train
num_bytes: 1369660519
num_examples: 200
download_size: 1057169221
dataset_size: 1369660519
configs:
- config_name: default
data_files:
- split: train
path: data/train-*