TNE-AI/saber-wa-chat-hi

这是一个由`gpt-4o`生成的合成印式英语（Hinglish）对话数据集，包含至少8轮的零售商和销售人员之间的对话。数据集以特定的温度`0.7`和`top-n`参数`0.6`生成。数据集的特征包括每个消息的标识符、角色、内容和用户名，以及数据集的训练分割信息，其中训练分割包含4050个示例和4595747字节的内存占用。

This is a synthetic Hinglish conversation dataset generated by `gpt-4o` with a temperature of `0.7` and a `top-n` parameter of `0.6`, containing at least 8 turns of dialogue between a retailer and a salesperson. The dataset features include an identifier, role, content, and username for each message, as well as information about the training split, which contains 4050 examples and occupies 4595747 bytes of memory.