orpo-dpo-mix-40k

Name: orpo-dpo-mix-40k
Creator: keyjjjj
Published: 2025-08-28 18:20:19
License: 暂无描述

AI_Studio2025-08-28 更新2025-05-31 收录

下载链接：

https://aistudio.baidu.com/datasetdetail/327398

下载链接

链接失效反馈

官方服务：

资源简介：

混合多个DPO数据集的ORPO训练专用偏好数据

Dedicated preference data for ORPO training, mixed from multiple DPO datasets

提供机构：

keyjjjj

orpo-dpo-mix-40k

# ORPO-DPO-mix-40k v1.2 ![image/webp](https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/s3uIwTgVl1sTm5_AX3rXH.webp) This dataset is designed for [ORPO](https://huggingf

魔搭社区2026-05-20 更新90

orpo-dpo-mix-40k-mlx

这是一个为MLX-LM优化的orpo-dpo-mix-40k数据集的拆分版本，专门用于ORPO训练。数据集被分为训练集（90%）、验证集（6%）和测试集（4%）。

Hugging Face2025-03-01 更新90

cleaned-orpo-dpo-mix-40k

该数据集包含多个特征，如'source', 'chosen', 'rejected', 'prompt', 'question'和'id'。其中'chosen'和'rejected'是列表类型，包含'content'和'role'两个子特征。数据集分为训练集，包含43704个样本，总大小为239288280字节。

Hugging Face2024-08-15 更新10

mlabonne/orpo-dpo-mix-40k

ORPO-DPO-mix-40k数据集专为ORPO或DPO训练设计，结合了多个高质量的DPO数据集，包括`argilla/Capybara-Preferences`、`argilla/distilabel-intel-orca-dpo-pairs`、`argilla/ultrafeedback-binarized-preferences-cleaned`、`argilla/distilabel-m

Hugging Face2024-06-07 更新500

orpo-dpo-mix-40k-flat

# ORPO-DPO-mix-40k-flat ![image/webp](https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/s3uIwTgVl1sTm5_AX3rXH.webp) This dataset is designed for [ORPO](https://huggingf

魔搭社区2025-11-27 更新30