From 18cbf8561d6c3fdceac47991ed16d35471823187 Mon Sep 17 00:00:00 2001
From: hiyouga <467089858@qq.com>
Date: Sat, 18 May 2024 21:15:20 +0800
Subject: [PATCH] update data readme

---
 data/README.md    | 191 +++++++++++++++++++++++++++++++++++++---------
 data/README_zh.md | 187 ++++++++++++++++++++++++++++++++++++++-------
 2 files changed, 315 insertions(+), 63 deletions(-)

diff --git a/data/README.md b/data/README.md
index b1368d4a..a467fe67 100644
--- a/data/README.md
+++ b/data/README.md
@@ -1,16 +1,17 @@
-If you are using a custom dataset, please add your **dataset description** to `dataset_info.json` according to the following format. We also provide several examples in the next section.
+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.
+
+Currently we support datasets in **alpaca** and **sharegpt** format.
 
 ```json
 "dataset_name": {
   "hf_hub_url": "the name of the dataset repository on the Hugging Face hub. (if specified, ignore script_url and file_name)",
-  "ms_hub_url": "the name of the dataset repository on the ModelScope hub. (if specified, ignore script_url and file_name)",
+  "ms_hub_url": "the name of the dataset repository on the Model Scope hub. (if specified, ignore script_url and file_name)",
   "script_url": "the name of the directory containing a dataset loading script. (if specified, ignore file_name)",
-  "file_name": "the name of the dataset file in this directory. (required if above are not specified)",
-  "file_sha1": "the SHA-1 hash value of the dataset file. (optional, does not affect training)",
+  "file_name": "the name of the dataset folder or dataset file in this directory. (required if above are not specified)",
+  "formatting": "the format of the dataset. (optional, default: alpaca, can be chosen from {alpaca, sharegpt})",
+  "ranking": "whether the dataset is a preference dataset or not. (default: False)",
   "subset": "the name of the subset. (optional, default: None)",
   "folder": "the name of the folder of the dataset repository on the Hugging Face hub. (optional, default: None)",
-  "ranking": "whether the dataset is a preference dataset or not. (default: false)",
-  "formatting": "the format of the dataset. (optional, default: alpaca, can be chosen from {alpaca, sharegpt})",
   "columns (optional)": {
     "prompt": "the column name in the dataset containing the prompts. (default: instruction)",
     "query": "the column name in the dataset containing the queries. (default: input)",
@@ -36,11 +37,15 @@ If you are using a custom dataset, please add your **dataset description** to `d
 }
 ```
 
-After that, you can load the custom dataset by specifying `--dataset dataset_name`.
+## Alpaca Format
 
-----
+### Supervised Fine-Tuning Dataset
 
-Currently we support dataset in **alpaca** or **sharegpt** format, the dataset in alpaca format should follow the below format:
+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.
 
 ```json
 [
@@ -57,7 +62,7 @@ Currently we support dataset in **alpaca** or **sharegpt** format, the dataset i
 ]
 ```
 
-Regarding the above dataset, the description in `dataset_info.json` should be:
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
 
 ```json
 "dataset_name": {
@@ -72,11 +77,9 @@ Regarding the above dataset, the description in `dataset_info.json` should be:
 }
 ```
 
-The `query` column will be concatenated with the `prompt` column and used as the human prompt, then the human prompt would be `prompt\nquery`. The `response` column represents the model response.
+### Pre-training Dataset
 
-The `system` column will be used as the system prompt. The `history` column is a list consisting string tuples representing prompt-response pairs in the history. Note that the responses in the history **will also be used for training** in supervised fine-tuning.
-
-For the **pre-training datasets**, only the `prompt` column will be used for training, for example:
+In pre-training, only the `prompt` column will be used for model learning.
 
 ```json
 [
@@ -85,7 +88,7 @@ For the **pre-training datasets**, only the `prompt` column will be used for tra
 ]
 ```
 
-Regarding the above dataset, the description in `dataset_info.json` should be:
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
 
 ```json
 "dataset_name": {
@@ -96,20 +99,24 @@ Regarding the above dataset, the description in `dataset_info.json` should be:
 }
 ```
 
-For the **preference datasets**, the `response` column should be a string list whose length is 2, with the preferred answers appearing first, for example:
+### Preference Dataset
+
+Preference datasets are used for reward modeling, DPO training and ORPO training.
+
+It requires a better response in `chosen` column and a worse response in `rejected` column.
 
 ```json
 [
   {
-    "instruction": "human instruction",
-    "input": "human input",
-    "chosen": "chosen answer",
-    "rejected": "rejected answer"
+    "instruction": "human instruction (required)",
+    "input": "human input (optional)",
+    "chosen": "chosen answer (required)",
+    "rejected": "rejected answer (required)"
   }
 ]
 ```
 
-Regarding the above dataset, the description in `dataset_info.json` should be:
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
 
 ```json
 "dataset_name": {
@@ -124,14 +131,86 @@ Regarding the above dataset, the description in `dataset_info.json` should be:
 }
 ```
 
-----
+### KTO Dataset
 
-The dataset in **sharegpt** format should follow the below format:
+KTO datasets require a extra `kto_tag` column containing the boolean human feedback.
+
+```json
+[
+  {
+    "instruction": "human instruction (required)",
+    "input": "human input (optional)",
+    "output": "model response (required)",
+    "kto_tag": "human feedback [true/false] (required)"
+  }
+]
+```
+
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
+
+```json
+"dataset_name": {
+  "file_name": "data.json",
+  "columns": {
+    "prompt": "instruction",
+    "query": "input",
+    "response": "output",
+    "kto_tag": "kto_tag"
+  }
+}
+```
+
+### Multimodal Dataset
+
+Multimodal datasets require a `images` column containing the paths to the input image. Currently we only support one image.
+
+```json
+[
+  {
+    "instruction": "human instruction (required)",
+    "input": "human input (optional)",
+    "output": "model response (required)",
+    "images": [
+      "image path (required)"
+    ]
+  }
+]
+```
+
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
+
+```json
+"dataset_name": {
+  "file_name": "data.json",
+  "columns": {
+    "prompt": "instruction",
+    "query": "input",
+    "response": "output",
+    "images": "images"
+  }
+}
+```
+
+## Sharegpt Format
+
+### 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.
+
+Note that the human and observation should appear in odd positions, while gpt and function should appear in even positions.
 
 ```json
 [
   {
     "conversations": [
+      {
+        "from": "human",
+        "value": "human instruction"
+      },
+      {
+        "from": "gpt",
+        "value": "model response"
+      },
       {
         "from": "human",
         "value": "human instruction"
@@ -147,7 +226,7 @@ The dataset in **sharegpt** format should follow the below format:
 ]
 ```
 
-Regarding the above dataset, the description in `dataset_info.json` should be:
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
 
 ```json
 "dataset_name": {
@@ -157,19 +236,61 @@ Regarding the above dataset, the description in `dataset_info.json` should be:
     "messages": "conversations",
     "system": "system",
     "tools": "tools"
-  },
-  "tags": {
-    "role_tag": "from",
-    "content_tag": "value",
-    "user_tag": "human",
-    "assistant_tag": "gpt"
   }
 }
 ```
 
-where the `messages` column should be a list following the `u/a/u/a/u/a` order.
+### Preference Dataset
 
-We also supports the dataset in the **openai** format:
+Preference datasets in sharegpt format also require a better message in `chosen` column and a worse message in `rejected` column.
+
+```json
+[
+  {
+    "conversations": [
+      {
+        "from": "human",
+        "value": "human instruction"
+      },
+      {
+        "from": "gpt",
+        "value": "model response"
+      },
+      {
+        "from": "human",
+        "value": "human instruction"
+      }
+    ],
+    "chosen": {
+      "from": "gpt",
+      "value": "chosen answer (required)"
+    },
+    "rejected": {
+      "from": "gpt",
+      "value": "rejected answer (required)"
+    }
+  }
+]
+```
+
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
+
+```json
+"dataset_name": {
+  "file_name": "data.json",
+  "formatting": "sharegpt",
+  "ranking": true,
+  "columns": {
+    "messages": "conversations",
+    "chosen": "chosen",
+    "rejected": "rejected"
+  }
+}
+```
+
+### OpenAI Format
+
+The openai format is simply a special case of the sharegpt format, where the first message may be a system prompt.
 
 ```json
 [
@@ -192,7 +313,7 @@ We also supports the dataset in the **openai** format:
 ]
 ```
 
-Regarding the above dataset, the description in `dataset_info.json` should be:
+Regarding the above dataset, the *dataset description* in `dataset_info.json` should be:
 
 ```json
 "dataset_name": {
@@ -211,4 +332,6 @@ Regarding the above dataset, the description in `dataset_info.json` should be:
 }
 ```
 
-Pre-training datasets and preference datasets are **incompatible** with the sharegpt format yet.
+The KTO datasets and multimodal datasets in sharegpt format are similar to the alpaca format.
+
+Pre-training datasets are **incompatible** with the sharegpt format.
diff --git a/data/README_zh.md b/data/README_zh.md
index deed94c5..61d60312 100644
--- a/data/README_zh.md
+++ b/data/README_zh.md
@@ -1,4 +1,6 @@
-如果您使用自定义数据集，请务必按照以下格式在 `dataset_info.json` 文件中添加**数据集描述**。我们在下面也提供了一些例子。
+`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。
 
 ```json
 "数据集名称": {
@@ -6,11 +8,10 @@
   "ms_hub_url": "ModelScope 的数据集仓库地址（若指定，则忽略 script_url 和 file_name）",
   "script_url": "包含数据加载脚本的本地文件夹名称（若指定，则忽略 file_name）",
   "file_name": "该目录下数据集文件的名称（若上述参数未指定，则此项必需）",
-  "file_sha1": "数据集文件的 SHA-1 哈希值（可选，留空不影响训练）",
+  "formatting": "数据集格式（可选，默认：alpaca，可以为 alpaca 或 sharegpt）",
+  "ranking": "是否为偏好数据集（可选，默认：False）",
   "subset": "数据集子集的名称（可选，默认：None）",
   "folder": "Hugging Face 仓库的文件夹名称（可选，默认：None）",
-  "ranking": "是否为偏好数据集（可选，默认：False）",
-  "formatting": "数据集格式（可选，默认：alpaca，可以为 alpaca 或 sharegpt）",
   "columns（可选）": {
     "prompt": "数据集代表提示词的表头名称（默认：instruction）",
     "query": "数据集代表请求的表头名称（默认：input）",
@@ -20,8 +21,8 @@
     "system": "数据集代表系统提示的表头名称（默认：None）",
     "tools": "数据集代表工具描述的表头名称（默认：None）",
     "images": "数据集代表图像输入的表头名称（默认：None）",
-    "chosen": "数据集代表更优回复的表头名称（默认：None）",
-    "rejected": "数据集代表更差回复的表头名称（默认：None）",
+    "chosen": "数据集代表更优回答的表头名称（默认：None）",
+    "rejected": "数据集代表更差回答的表头名称（默认：None）",
     "kto_tag": "数据集代表 KTO 标签的表头名称（默认：None）"
   },
   "tags（可选，用于 sharegpt 格式）": {
@@ -31,16 +32,20 @@
     "assistant_tag": "消息中代表助手的 role_tag（默认：gpt）",
     "observation_tag": "消息中代表工具返回结果的 role_tag（默认：observation）",
     "function_tag": "消息中代表工具调用的 role_tag（默认：function_call）",
-    "system_tag": "消息中代表系统提示的 role_tag（默认：system，会覆盖 system 列）"
+    "system_tag": "消息中代表系统提示的 role_tag（默认：system，会覆盖 system column）"
   }
 }
 ```
 
-然后，可通过使用 `--dataset 数据集名称` 参数加载自定义数据集。
+## Alpaca 格式
 
-----
+### 指令监督微调数据集
 
-该项目目前支持两种格式的数据集：**alpaca** 和 **sharegpt**，其中 alpaca 格式的数据集按照以下方式组织：
+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。
+
+`history` 列是由多个字符串二元组构成的列表，分别代表历史消息中每轮对话的指令和回答。注意在指令监督微调时，历史消息中的回答内容**也会被用于模型学习**。
 
 ```json
 [
@@ -57,7 +62,7 @@
 ]
 ```
 
-对于上述格式的数据，`dataset_info.json` 中的描述应为：
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
 
 ```json
 "数据集名称": {
@@ -72,11 +77,9 @@
 }
 ```
 
-其中 `query` 列对应的内容会与 `prompt` 列对应的内容拼接后作为人类指令，即人类指令为 `prompt\nquery`。`response` 列对应的内容为模型回答。
+### 预训练数据集
 
-`system` 列对应的内容将被作为系统提示词。`history` 列是由多个字符串二元组构成的列表，分别代表历史消息中每轮的指令和回答。注意在指令监督学习时，历史消息中的回答**也会被用于训练**。
-
-对于**预训练数据集**，仅 `prompt` 列中的内容会用于模型训练，例如：
+对于**预训练数据集**，仅 `prompt` 列中的内容会用于模型学习，例如：
 
 ```json
 [
@@ -85,7 +88,7 @@
 ]
 ```
 
-对于上述格式的数据，`dataset_info.json` 中的描述应为：
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
 
 ```json
 "数据集名称": {
@@ -96,20 +99,24 @@
 }
 ```
 
-对于**偏好数据集**，`response` 列应当是一个长度为 2 的字符串列表，排在前面的代表更优的回答，例如：
+### 偏好数据集
+
+偏好数据集用于奖励模型训练、DPO 训练和 ORPO 训练。
+
+它需要在 `chosen` 列中提供更优的回答，并在 `rejected` 列中提供更差的回答。
 
 ```json
 [
   {
-    "instruction": "人类指令",
-    "input": "人类输入",
-    "chosen": "优质回答",
-    "rejected": "劣质回答"
+    "instruction": "人类指令（必填）",
+    "input": "人类输入（选填）",
+    "chosen": "优质回答（必填）",
+    "rejected": "劣质回答（必填）"
   }
 ]
 ```
 
-对于上述格式的数据，`dataset_info.json` 中的描述应为：
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
 
 ```json
 "数据集名称": {
@@ -124,14 +131,86 @@
 }
 ```
 
-----
+### KTO 数据集
 
-而 **sharegpt** 格式的数据集按照以下方式组织：
+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
+[
+  {
+    "instruction": "人类指令（必填）",
+    "input": "人类输入（选填）",
+    "output": "模型回答（必填）",
+    "kto_tag": "人类反馈 [true/false]（必填）"
+  }
+]
+```
+
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
+
+```json
+"数据集名称": {
+  "file_name": "data.json",
+  "columns": {
+    "prompt": "instruction",
+    "query": "input",
+    "response": "output",
+    "kto_tag": "kto_tag"
+  }
+}
+```
+
+### 多模态数据集
+
+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
+[
+  {
+    "instruction": "人类指令（必填）",
+    "input": "人类输入（选填）",
+    "output": "模型回答（必填）",
+    "images": [
+      "图像路径（必填）"
+    ]
+  }
+]
+```
+
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
+
+```json
+"数据集名称": {
+  "file_name": "data.json",
+  "columns": {
+    "prompt": "instruction",
+    "query": "input",
+    "response": "output",
+    "images": "images"
+  }
+}
+```
+
+## Sharegpt 格式
+
+### 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。
+
+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
 [
   {
     "conversations": [
+      {
+        "from": "human",
+        "value": "人类指令"
+      },
+      {
+        "from": "gpt",
+        "value": "模型回答"
+      },
       {
         "from": "human",
         "value": "人类指令"
@@ -147,7 +226,7 @@
 ]
 ```
 
-对于上述格式的数据，`dataset_info.json` 中的描述应为：
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
 
 ```json
 "数据集名称": {
@@ -167,9 +246,57 @@
 }
 ```
 
-其中 `messages` 列应当是一个列表，且符合 `人类/模型/人类/模型/人类/模型` 的顺序。
+### 偏好数据集
 
-我们同样支持 **openai** 格式的数据集：
+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
+[
+  {
+    "conversations": [
+      {
+        "from": "human",
+        "value": "人类指令"
+      },
+      {
+        "from": "gpt",
+        "value": "模型回答"
+      },
+      {
+        "from": "human",
+        "value": "人类指令"
+      }
+    ],
+    "chosen": {
+      "from": "gpt",
+      "value": "优质回答"
+    },
+    "rejected": {
+      "from": "gpt",
+      "value": "劣质回答"
+    }
+  }
+]
+```
+
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
+
+```json
+"数据集名称": {
+  "file_name": "data.json",
+  "formatting": "sharegpt",
+  "ranking": true,
+  "columns": {
+    "messages": "conversations",
+    "chosen": "chosen",
+    "rejected": "rejected"
+  }
+}
+```
+
+### OpenAI 格式
+
+OpenAI 格式仅仅是 sharegpt 格式的一种特殊情况，其中第一条消息可能是系统提示词。
 
 ```json
 [
@@ -192,7 +319,7 @@
 ]
 ```
 
-对于上述格式的数据，`dataset_info.json` 中的描述应为：
+对于上述格式的数据，`dataset_info.json` 中的*数据集描述*应为：
 
 ```json
 "数据集名称": {
@@ -211,4 +338,6 @@
 }
 ```
 
-预训练数据集和偏好数据集**尚不支持** sharegpt 格式。
+Sharegpt 格式中的 KTO 数据集和多模态数据集与 alpaca 格式的类似。
+
+预训练数据集**不支持** sharegpt 格式。