HachiML/Evol-hh-rlhf-gen3-1k
收藏资源简介:
--- dataset_info: features: - name: No. dtype: int64 - name: seed_id dtype: int64 - name: generation dtype: int64 - name: evol_history sequence: string - name: instruction dtype: string - name: output dtype: string splits: - name: train num_bytes: 2340124 num_examples: 1017 download_size: 1098508 dataset_size: 2340124 configs: - config_name: default data_files: - split: train path: data/train-* license: mit task_categories: - text-generation language: - ja tags: - synthetic - evol-instruct size_categories: - 1K<n<10K --- # Evol-hh-rlhf-gen3-1k <!-- Provide a quick summary of the dataset. --> Evol-hh-rlhf-gen3-1kは、 - [kunishou/hh-rlhf-49k-ja](https://huggingface.co/datasets/kunishou/hh-rlhf-49k-ja)をseed tasksとして - [Evol-Instruction](https://arxiv.org/abs/2304.12244)の手法 - [mistralai/Mixtral-8x22B-Instruct-v0.1](https://huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1) で作った合成データ(Synthetic data)です。 モデルの利用には[Deepinfra](https://deepinfra.com/mistralai/Mixtral-8x22B-Instruct-v0.1/api?example=openai-python)を利用しています。 <!-- This dataset card aims to be a base template for new datasets. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/datasetcard_template.md?plain=1). --> ## Dataset Details ### Dataset Description <!-- Provide a longer summary of what this dataset is. --> - **Curated by:** [HachiML](https://huggingface.co/HachiML) - **Language(s) (NLP):** Japanese - **License:** Apache 2.0 - **Github:** [Evol-Instruct-jp](https://github.com/Hajime-Y/Evol-Instruct-jp) ## Uses <!-- Address questions around how the dataset is intended to be used. --> ```Python # library from datasets import load_dataset # Load dataset. dataset = load_dataset("HachiML/Evol-hh-rlhf-gen3-1k") ``` ## Code **Github:** [Evol-Instruct-jp](https://github.com/Hajime-Y/Evol-Instruct-jp) にコードを置いています。このコードを元に、以下の設定で生成しました。 ```Python !python main.py \ --input_file "./data/hh-rlhf-49k-ja.jsonl" \ --output_file "./output/generated.json" \ --eliminated_file "./output/eliminated.json" \ --model "mistralai/Mixtral-8x22B-Instruct-v0.1" \ --num_instructions_to_generate 1000 \ --subset_size 100 ``` 1k recordsの生成はsubset37 で目標数に達しました。続きから生成を開始する際は start_subset_index=38 の設定が必要です。
数据集概述
数据集名称
- 名称: Evol-hh-rlhf-gen3-1k
数据集特征
- 特征列表:
- No. (int64)
- seed_id (int64)
- generation (int64)
- evol_history (sequence: string)
- instruction (string)
- output (string)
数据集拆分
- 拆分详情:
- train:
- 示例数量: 1017
- 数据大小: 2340124字节
- train:
数据集大小
- 下载大小: 1098508字节
- 数据集大小: 2340124字节
配置信息
- 默认配置:
- 数据文件路径: data/train-*
许可证
- 许可证: MIT
任务类别
- 任务类别: text-generation
语言
- 语言: Japanese
标签
- 标签:
- synthetic
- evol-instruct
大小类别
- 大小类别: 1K<n<10K



