mesolitica/DPO-filtered-aya_dataset-zsm
收藏Hugging Face2024-02-13 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/mesolitica/DPO-filtered-aya_dataset-zsm
下载链接
链接失效反馈官方服务:
资源简介:
---
language:
- ms
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
splits:
- name: train
num_bytes: 12310875
num_examples: 10073
download_size: 5813801
dataset_size: 12310875
---
# DPO Binarized filtered-aya_dataset-zsm
DPO binarized style using filtered https://huggingface.co/datasets/CohereForAI/aya_dataset on `zsm` language only, after that we use https://huggingface.co/mesolitica/malaysian-mistral-7b-32k-instructions-v4 to generate the outputs and the generated outputs use `rejected` column.
Read more about DPO binarized style dataset at https://huggingface.co/docs/trl/main/en/dpo_trainer
提供机构:
mesolitica
原始信息汇总
DPO Binarized filtered-aya_dataset-zsm
数据集信息
特征
- prompt: 数据类型为字符串。
- chosen: 数据类型为字符串。
- rejected: 数据类型为字符串。
分割
- train:
- 字节数: 12310875
- 样本数: 10073
大小
- 下载大小: 5813801
- 数据集大小: 12310875



