DaftP/Home-Assistant-requests-for-intent-detection-and-function-recognition

--- license: mit task_categories: - question-answering - text-generation tags: - automation - home - assistant language: ["en", "es", "fr", "de", "pl"] pretty_name: Home Assistant Requests V2 size_categories: - 10K<n<100k --- # Home Assistant Requests V2 Dataset This dataset contains a list of requests and responses for a user interacting with a personal assistant that controls an instance of [Home Assistant](https://www.home-assistant.io/). The updated V2 of the dataset is now multilingual, containing data in English, German, French, Spanish, and Polish. The dataset also contains multiple "personalities" for the assistant to respond in, such as a formal assistant, a sarcastic assistant, and a friendly assistant. Lastly, the dataset has been updated to fully support modern tool-calling formats. > NOTE: If you are viewing this dataset on HuggingFace, you can download the "small" dataset variant directly from the "Files and versions" tab. ## Assembling the dataset The dataset is generated from the different CSV "piles". The "piles" contain different chunks of requests that are assembled into a final context that is presented to the LLM. For example, `piles/<language>/pile_of_device_names.csv` contains only names of various devices to be used as part of context as well as inserted into `piles/<language>/pile_of_templated_actions.csv` and `piles/<language>/pile_of_status_requests.csv`. The logic for assembling the final dataset from the piles is contained in [generate_data.py](./generate_data.py). ### Prepare environment Start by installing system dependencies: `sudo apt-get install python3-dev` Then create a Python virtual environment and install all necessary library: ``` python3 -m venv .generate_data source .generate_data/bin/activate pip3 install -r requirements.txt ``` ### Generating the dataset from piles `python3 generate_data.py --train --test --small --language english german french spanish polish` Supported dataset splits are `--test`, `--train`, & `--sample` Arguments to set the train dataset size are `--small`, `--medium`, `--large`, & `--xl`. Languages can be enabled using `--language english german french spanish polish` # Router, Engineer and Responder Datasets Those 3 datasets are the result of the formatting operation on the original `home_assistant_train.jsonl` file ### Router Dataset This dataset is generated in order to train a LLM to better detect intents from user's interactions in a smart home environment, specifically integrated with Home Assistant framework ### Engineer Dataset This dataset is generated in order to train a LLM to better generate commands from Home Assistant services and smart home context, starting from detected intent and target entity. ### Responder Dataset This dataset is generated in order to train a LLM to better generate responses for status reading queries about smart home context devices--- 许可证：MIT 任务类别： - 问答 - 文本生成 标签： - 自动化 - 家庭 - 助手 语言：["英语", "西班牙语", "法语", "德语", "波兰语"] 展示名称：家庭助手请求V2 规模类别：10K<n<100k --- # 家庭助手请求V2数据集（Home Assistant Requests V2 Dataset） 本数据集收录了用户与控制家庭助手（Home Assistant）实例的个人助手交互时产生的全部请求与响应内容。 本数据集的更新版V2现已支持多语言，涵盖英语、德语、法语、西班牙语与波兰语五类语言。此外，数据集内置了多种助手回复人设，包括正式型、戏谑型与友好型助手。本次更新还全面兼容现代工具调用格式。 > 注意：若您在HuggingFace平台查看本数据集，可直接从「文件与版本」标签页下载「小型」数据集变体。 ## 数据集构建流程 本数据集由多个CSV「语料包（piles）」拼接生成。这些「语料包」包含各类请求片段，经整合后形成供大语言模型（Large Language Model，LLM）使用的最终上下文。例如，`piles/<语言>/pile_of_device_names.csv`仅收录各类设备名称，既可作为上下文内容使用，也可插入至`piles/<语言>/pile_of_templated_actions.csv`与`piles/<语言>/pile_of_status_requests.csv`中。从语料包构建最终数据集的逻辑代码位于[generate_data.py](./generate_data.py)。 ### 环境准备 首先安装系统依赖项： `sudo apt-get install python3-dev` 随后创建Python虚拟环境并安装所有必需的依赖库： python3 -m venv .generate_data source .generate_data/bin/activate pip3 install -r requirements.txt ### 从语料包生成数据集 执行以下命令即可生成数据集： `python3 generate_data.py --train --test --small --language english german french spanish polish` 支持的数据集划分参数包括`--test`、`--train`与`--sample`；用于设置训练集规模的参数有`--small`（小型）、`--medium`（中型）、`--large`（大型）与`--xl`（超大型）；可通过`--language english german french spanish polish`参数启用对应语言的数据集。 ## 路由、工程与响应数据集 上述三类数据集均由原始`home_assistant_train.jsonl`文件经格式化处理后得到。 ### 路由数据集 本数据集用于训练大语言模型，使其能够更精准地识别智能家居场景下用户交互中的意图，且该场景已与Home Assistant框架深度集成。 ### 工程数据集 本数据集用于训练大语言模型，使其能够基于已识别的意图与目标实体，从Home Assistant服务与智能家居上下文中生成准确的控制命令。 ### 响应数据集 本数据集用于训练大语言模型，使其能够针对智能家居设备的状态查询生成更贴合场景的回复内容。