README.md

---
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](https://luma.com/we2jbbp3)で作成したモデルのCPTに利用したデータセットです。