metadata
dataset_info:
- config_name: emilia-en
features:
- name: input_ids
list: int64
- name: labels
list: int64
- name: attention_mask
list: int64
splits:
- name: train
num_bytes: 30603413592
num_examples: 2740520
download_size: 6650332901
dataset_size: 30603413592
- config_name: emilia-ja
features:
- name: input_ids
list: int64
- name: attention_mask
list: int64
- name: labels
list: int64
splits:
- name: train
num_bytes: 11847406500
num_examples: 1285235
download_size: 2670616914
dataset_size: 11847406500
- config_name: wikipedia-en
features:
- name: input_ids
list: int64
- name: labels
list: int64
- name: attention_mask
list: int64
splits:
- name: train
num_bytes: 25507946208
num_examples: 1500000
download_size: 4347705671
dataset_size: 25507946208
- config_name: wikipedia-ja
features:
- name: input_ids
list: int64
- name: labels
list: int64
- name: attention_mask
list: int64
splits:
- name: train
num_bytes: 50593431852
num_examples: 1472757
download_size: 7328921641
dataset_size: 50593431852
configs:
- config_name: emilia-en
data_files:
- split: train
path: emilia-en/train-*
- config_name: emilia-ja
data_files:
- split: train
path: emilia-ja/train-*
- config_name: wikipedia-en
data_files:
- split: train
path: wikipedia-en/train-*
- config_name: wikipedia-ja
data_files:
- split: train
path: wikipedia-ja/train-*
license: cc-by-nc-4.0
task_categories:
- automatic-speech-recognition
- text-to-speech
- audio-to-audio
language:
- ja
- en
LiquidAI-Hackathon-Tokyo-CPT-Data
Liquid AI Hackathon Tokyoで作成したモデルのCPTに利用したデータセットです。