From ca48f90f1eb9828300635bdaee6c10d6cc632d3d Mon Sep 17 00:00:00 2001 From: hiyouga <467089858@qq.com> Date: Sat, 18 May 2024 21:37:38 +0800 Subject: [PATCH] update data readme --- data/README.md | 30 +++++++++++++++++++++--------- data/README_zh.md | 34 ++++++++++++++++++++-------------- 2 files changed, 41 insertions(+), 23 deletions(-) diff --git a/data/README.md b/data/README.md index a467fe67..3e96fbeb 100644 --- a/data/README.md +++ b/data/README.md @@ -1,4 +1,4 @@ -The `dataset_info.json` contains all available datasets. If you are using a custom dataset, please make sure to add a *dataset description* in `dataset_info.json` and specify `dataset: dataset_name` before training to use it. +The [dataset_info.json](dataset_info.json) contains all available datasets. If you are using a custom dataset, please **make sure** to add a *dataset description* in `dataset_info.json` and specify `dataset: dataset_name` before training to use it. Currently we support datasets in **alpaca** and **sharegpt** format. @@ -41,11 +41,13 @@ Currently we support datasets in **alpaca** and **sharegpt** format. ### Supervised Fine-Tuning Dataset +* [Example dataset](alpaca_en_demo.json) + In supervised fine-tuning, the `instruction` column will be concatenated with the `input` column and used as the human prompt, then the human prompt would be `instruction\ninput`. The `output` column represents the model response. The `system` column will be used as the system prompt if specified. -The `history` column is a list consisting string tuples representing prompt-response pairs in the history messages. Note that the responses in the history **will also be learned by the model** in supervised fine-tuning. +The `history` column is a list consisting of string tuples representing prompt-response pairs in the history messages. Note that the responses in the history **will also be learned by the model** in supervised fine-tuning. ```json [ @@ -79,7 +81,9 @@ Regarding the above dataset, the *dataset description* in `dataset_info.json` sh ### Pre-training Dataset -In pre-training, only the `prompt` column will be used for model learning. +- [Example dataset](c4_demo.json) + +In pre-training, only the `text` column will be used for model learning. ```json [ @@ -133,6 +137,8 @@ Regarding the above dataset, the *dataset description* in `dataset_info.json` sh ### KTO Dataset +- [Example dataset](kto_en_demo.json) + KTO datasets require a extra `kto_tag` column containing the boolean human feedback. ```json @@ -162,7 +168,9 @@ Regarding the above dataset, the *dataset description* in `dataset_info.json` sh ### Multimodal Dataset -Multimodal datasets require a `images` column containing the paths to the input image. Currently we only support one image. +- [Example dataset](mllm_demo.json) + +Multimodal datasets require a `images` column containing the paths to the input images. Currently we only support one image. ```json [ @@ -195,7 +203,9 @@ Regarding the above dataset, the *dataset description* in `dataset_info.json` sh ### Supervised Fine-Tuning Dataset -Compared to the alpaca format, the sharegpt format allows the datasets have more **roles**, such as human, gpt, observation and function. They are presented in a list of objects in the `conversations` column. +- [Example dataset](glaive_toolcall_en_demo.json) + +Compared to the alpaca format, the sharegpt format allows the datasets have **more roles**, such as human, gpt, observation and function. They are presented in a list of objects in the `conversations` column. Note that the human and observation should appear in odd positions, while gpt and function should appear in even positions. @@ -208,12 +218,12 @@ Note that the human and observation should appear in odd positions, while gpt an "value": "human instruction" }, { - "from": "gpt", - "value": "model response" + "from": "function_call", + "value": "tool arguments" }, { - "from": "human", - "value": "human instruction" + "from": "observation", + "value": "tool result" }, { "from": "gpt", @@ -242,6 +252,8 @@ Regarding the above dataset, the *dataset description* in `dataset_info.json` sh ### Preference Dataset +- [Example dataset](dpo_en_demo.json) + Preference datasets in sharegpt format also require a better message in `chosen` column and a worse message in `rejected` column. ```json diff --git a/data/README_zh.md b/data/README_zh.md index 61d60312..d8a2419e 100644 --- a/data/README_zh.md +++ b/data/README_zh.md @@ -1,4 +1,4 @@ -`dataset_info.json` 包含了所有可用的数据集。如果您希望使用自定义数据集，请务必在 `dataset_info.json` 文件中添加*数据集描述*，并通过修改 `dataset: 数据集名称` 配置来使用数据集。 +[dataset_info.json](dataset_info.json) 包含了所有可用的数据集。如果您希望使用自定义数据集，请**务必**在 `dataset_info.json` 文件中添加*数据集描述*，并通过修改 `dataset: 数据集名称` 配置来使用数据集。 目前我们支持 **alpaca** 格式和 **sharegpt** 格式的数据集。 @@ -41,6 +41,8 @@ ### 指令监督微调数据集 +- [样例数据集](alpaca_zh_demo.json) + 在指令监督微调时，`instruction` 列对应的内容会与 `input` 列对应的内容拼接后作为人类指令，即人类指令为 `instruction\ninput`。而 `output` 列对应的内容为模型回答。 如果指定，`system` 列对应的内容将被作为系统提示词。 @@ -79,7 +81,9 @@ ### 预训练数据集 -对于**预训练数据集**，仅 `prompt` 列中的内容会用于模型学习，例如： +- [样例数据集](c4_demo.json) + +在预训练时，只有 `text` 列中的内容会用于模型学习。 ```json [ @@ -133,6 +137,8 @@ ### KTO 数据集 +- [样例数据集](kto_en_demo.json) + KTO 数据集需要额外添加一个 `kto_tag` 列，包含 bool 类型的人类反馈。 ```json @@ -162,6 +168,8 @@ KTO 数据集需要额外添加一个 `kto_tag` 列，包含 bool 类型的人 ### 多模态数据集 +- [样例数据集](mllm_demo.json) + 多模态数据集需要额外添加一个 `images` 列，包含输入图像的路径。目前我们仅支持单张图像输入。 ```json @@ -195,9 +203,11 @@ KTO 数据集需要额外添加一个 `kto_tag` 列，包含 bool 类型的人 ### 指令监督微调数据集 -相比 alpaca 格式的数据集，sharegpt 格式支持更多的**角色种类**，例如 human、gpt、observation、function 等等。它们构成一个对象列表呈现在 `conversations` 列中。 +- [样例数据集](glaive_toolcall_zh_demo.json) -其中 human 和 observation 必须出现在奇数位置，gpt 和 function 必须出现在偶数位置。 +相比 alpaca 格式的数据集，sharegpt 格式支持**更多的角色种类**，例如 human、gpt、observation、function 等等。它们构成一个对象列表呈现在 `conversations` 列中。 + +注意其中 human 和 observation 必须出现在奇数位置，gpt 和 function 必须出现在偶数位置。 ```json [ @@ -208,12 +218,12 @@ KTO 数据集需要额外添加一个 `kto_tag` 列，包含 bool 类型的人 "value": "人类指令" }, { - "from": "gpt", - "value": "模型回答" + "from": "function_call", + "value": "工具参数" }, { - "from": "human", - "value": "人类指令" + "from": "observation", + "value": "工具结果" }, { "from": "gpt", @@ -236,18 +246,14 @@ KTO 数据集需要额外添加一个 `kto_tag` 列，包含 bool 类型的人 "messages": "conversations", "system": "system", "tools": "tools" - }, - "tags": { - "role_tag": "from", - "content_tag": "value", - "user_tag": "human", - "assistant_tag": "gpt" } } ``` ### 偏好数据集 +- [样例数据集](dpo_zh_demo.json) + Sharegpt 格式的偏好数据集同样需要在 `chosen` 列中提供更优的消息，并在 `rejected` 列中提供更差的消息。 ```json