metadata
dataset_info:
features:
- name: session_id
dtype: string
- name: turn_index
dtype: int64
- name: audio_file
dtype: string
- name: transcript
dtype: string
- name: chat_history_last_9
dtype: string
- name: chat_history_last_9_tokenized
sequence: int64
- name: speaker_text_last_9_tokenized
sequence: int64
- name: attention_mask
sequence: int64
- name: label_semantics
dtype: string
- name: label
dtype: string
splits:
- name: train
num_bytes: 196537438
num_examples: 22266
download_size: 3094864
dataset_size: 196537438
configs:
- config_name: default
data_files:
- split: train
path: data/train-*