exp_rpt_stack-junit-v6-qwen3.5-122b-131k-opencode-sft-serveparity
收藏资源简介:
该数据集是一个面向多轮对话与工具调用任务的结构化数据集,包含660个训练样本。每个样本由对话消息序列(messages)组成,其中每条消息包括角色(如用户或助手)、文本内容以及可选的工具调用信息(如工具类型和函数参数);同时提供工具定义(tools)描述可调用的工具,任务类型(task)标识对话所属类别,以及统计指标如对话轮数(num_turns)和工具调用次数(num_tool_calls)。该数据集适用于训练或评估支持工具调用的对话系统,例如基于大语言模型的智能助手或任务导向型对话代理。
This dataset is a structured dataset for multi-turn dialogue and tool calling tasks, containing 660 training samples. Each sample consists of a dialogue message sequence (messages), where each message includes a role (e.g., user or assistant), text content, and optional tool calling information (e.g., tool type and function parameters); it also provides tool definitions (tools) describing callable tools, task type (task) identifying the dialogue category, and statistical metrics such as the number of dialogue turns (num_turns) and tool calls (num_tool_calls). This dataset is suitable for training or evaluating dialogue systems that support tool calling, such as large language model-based intelligent assistants or task-oriented dialogue agents.




