five

sapienzanlp/it-Magpie-Llama-3.1-Pro-300K-Filtered-easy

收藏
Hugging Face2024-12-05 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/sapienzanlp/it-Magpie-Llama-3.1-Pro-300K-Filtered-easy
下载链接
链接失效反馈
官方服务:
资源简介:
--- language: - en - it size_categories: - 10K<n<100K dataset_info: features: - name: id dtype: int64 - name: text dtype: string - name: uuid dtype: string - name: model dtype: string - name: gen_input_configs struct: - name: input_generator dtype: string - name: pre_query_template dtype: string - name: seed dtype: 'null' - name: temperature dtype: float64 - name: top_p dtype: float64 - name: instruction dtype: string - name: response dtype: string - name: task_category dtype: string - name: other_task_category sequence: string - name: task_category_generator dtype: string - name: difficulty dtype: string - name: intent dtype: string - name: knowledge dtype: string - name: difficulty_generator dtype: string - name: input_quality dtype: string - name: quality_explanation dtype: string - name: quality_generator dtype: string - name: llama_guard_2 dtype: string - name: reward_model dtype: string - name: instruct_reward dtype: float64 - name: min_neighbor_distance dtype: float64 - name: repeat_count dtype: int64 - name: min_similar_uuid dtype: string - name: instruction_length dtype: int64 - name: response_length dtype: int64 - name: language dtype: string - name: instruction_it dtype: string - name: response_it dtype: string - name: messages list: - name: content dtype: string - name: role dtype: string splits: - name: train num_bytes: 571262153 num_examples: 59070 download_size: 266524472 dataset_size: 571262153 configs: - config_name: default data_files: - split: train path: data/train-* --- # Dataset Card: Magpie-Llama-3.1-Pro-300K-Filtered (Italian Translation) ## Dataset Summary This dataset is the **Italian translation** of the [Magpie-Llama-3.1-Pro-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-300K-Filtered) dataset, designed specifically for **instruction tuning** in Italian. We translated the instances whose 'difficulty' value is 'easy'. The translation was carried out using **TowerInstruct-Mistral-7B-v0.2**. - **Languages**: Italian (translated from English) - **Purpose**: Instruction tuning in Italian - **Train Size**: 59070 New fields added include: ```python { "messages": list # the translated conversation in Italian, presented in chat template format. "instruction_it": str # the translated instruction in Italian. "response_it_": str # the translated response in Italian. } ``` These additions make the dataset suitable for Italian-speaking AI models and multilingual systems requiring instruction following data in Italian. ## Author Alessandro Scirè (scire@babelscape.com)

语言: - 英语 - 意大利语 规模类别: - 10000条 < 样本数量 < 100000条 数据集信息: 特征: - 名称:id,数据类型:int64 - 名称:text,数据类型:字符串 - 名称:uuid(通用唯一识别码,UUID),数据类型:字符串 - 名称:model,数据类型:字符串 - 名称:gen_input_configs(生成输入配置),结构体: - 名称:input_generator,数据类型:字符串 - 名称:pre_query_template(前置查询模板),数据类型:字符串 - 名称:seed(随机种子),数据类型:空值 - 名称:temperature(温度系数),数据类型:float64 - 名称:top_p(核采样阈值),数据类型:float64 - 名称:instruction(指令),数据类型:字符串 - 名称:response(回复),数据类型:字符串 - 名称:task_category(任务类别),数据类型:字符串 - 名称:other_task_category(其他任务类别),字符串序列 - 名称:task_category_generator(任务类别生成器),数据类型:字符串 - 名称:difficulty(难度等级),数据类型:字符串 - 名称:intent(意图),数据类型:字符串 - 名称:knowledge(知识范围),数据类型:字符串 - 名称:difficulty_generator(难度生成器),数据类型:字符串 - 名称:input_quality(输入质量),数据类型:字符串 - 名称:quality_explanation(质量说明),数据类型:字符串 - 名称:quality_generator(质量生成器),数据类型:字符串 - 名称:Llama Guard 2,数据类型:字符串 - 名称:reward_model(奖励模型),数据类型:字符串 - 名称:instruct_reward(指令奖励分数),数据类型:float64 - 名称:min_neighbor_distance(最小邻域距离),数据类型:float64 - 名称:repeat_count(重复次数),数据类型:int64 - 名称:min_similar_uuid(最相似样本的UUID),数据类型:字符串 - 名称:instruction_length(指令长度),数据类型:int64 - 名称:response_length(回复长度),数据类型:int64 - 名称:language(语言),数据类型:字符串 - 名称:instruction_it(意大利语指令),数据类型:字符串 - 名称:response_it(意大利语回复),数据类型:字符串 - 名称:messages(对话消息),列表: - 名称:content(消息内容),数据类型:字符串 - 名称:role(角色),数据类型:字符串 划分方式: - 名称:train(训练集),字节数:571262153,样本数:59070 下载大小:266524472字节 数据集总大小:571262153字节 配置项: - 配置名称:default(默认),数据文件: - 划分集:train,路径:data/train-* # 数据集卡片:Magpie-Llama-3.1-Pro-300K-Filtered(意大利语翻译版) ## 数据集概览 本数据集为[Magpie-Llama-3.1-Pro-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-300K-Filtered)数据集的**意大利语翻译版本**,专为意大利语场景下的**指令微调**设计。我们仅翻译了难度等级为「简单」的样本,翻译工作由**TowerInstruct-Mistral-7B-v0.2**完成。 - **语言**:意大利语(由英语翻译而来) - **用途**:意大利语场景下的指令微调 - **训练集规模**:59070 新增字段如下: python { "messages": list # 采用聊天模板格式组织的意大利语翻译对话 "instruction_it": str # 意大利语翻译后的指令 "response_it_": str # 意大利语翻译后的回复 } 上述新增字段使本数据集适配意大利语AI模型,以及需要意大利语指令跟随数据的多语言系统。 ## 作者 Alessandro Scirè (scire@babelscape.com)
提供机构:
sapienzanlp
二维码
社区交流群
二维码
科研交流群
商业服务