dariolopez/Llama-2-oasst1-es

---
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: text
dtype: string
splits:
- name: train
num_bytes: 4524060
num_examples: 3909
download_size: 2528456
dataset_size: 4524060
license: apache-2.0
language:
- es
size_categories:
- 1K<n<10K
---
# OpenAssistant Conversations Spanish Dataset (OASST1-es) for Llama-2

## Dataset Summary

Subset of the original [OpenAssistant Conversations Dataset (OASST)](https://huggingface.co/datasets/OpenAssistant/oasst1).

* Filtered by `lang=es`.
* Formatted according to the Llama-2 pattern: "\<s> [INST] user prompt [/INST] output model \</s>"
* Select the best ranked output (Some instructions had multiple outputs ranked by humans).
* Select only the first level of the tree conversation.

## Dataset Structure

The dataset has 3909 rows of tuples (instructions and outputs).

原始信息汇总

OpenAssistant Conversations Spanish Dataset (OASST1-es) for Llama-2

数据集概述

• 该数据集是原始OpenAssistant Conversations Dataset (OASST)的子集。
• 过滤条件为lang=es，即西班牙语。
• 格式遵循Llama-2模式："<s> [INST] user prompt [/INST] output model </s>"。
• 选择最佳排名的输出（某些指令有多个输出，由人工排名）。
• 仅选择树状对话的第一层级。

数据集结构

• 数据集包含3909行，每行是一个元组（指令和输出）。