Yehor/kulyk-bidirectional

该数据集是一个大规模对话数据集，包含81810830个训练示例和测试分割。每个示例由消息列表组成，每条消息包括content（文本内容）和role（角色，如用户或助手）字段，适用于自然语言处理和人工智能对话模型的训练与评估。数据集总大小约为27.8 GB，下载大小约为9.4 GB。

This dataset is a large-scale conversational dataset containing 81,810,830 training examples and a test split. Each example consists of a list of messages, with each message including content (text content) and role (e.g., user or assistant) fields, suitable for training and evaluating natural language processing and AI dialogue models. The total dataset size is approximately 27.8 GB, with a download size of approximately 9.4 GB.