INFINITY1023/MultilingualDriverCommands

多语言驾驶员命令数据集是一个合成文本数据集，用于训练和评估车内语音助手的意图分类模型。该数据集包含俄语和英语的短驾驶员命令。每个示例代表用户对车辆语音助手的短语，并映射到一个意图类别。该数据集专为自然语言理解任务设计，特别是多类别意图分类。数据集通过合成方式生成，经过手动验证，在类别之间保持平衡，并丰富了额外的罕见驾驶相关场景。Multilingual Driver Commands Dataset is a synthetic text dataset for training and evaluating intent classification models for an in-car voice assistant. The dataset contains short driver commands in Russian and English. Each example represents a user phrase addressed to a vehicle voice assistant and is mapped to an intent class. The dataset is designed for Natural Language Understanding (NLU) tasks, especially multiclass intent classification. The dataset was generated synthetically, manually validated, balanced across classes, and enriched with additional rare driving-related scenarios.