exp_rpt_stack-pytest-large-qwen3.5-122b-131k-opencode-sft-serveparity
收藏资源简介:
该数据集是一个面向对话与工具调用任务的结构化数据集,包含1328个训练样本。每个样本包含以下核心字段:messages字段记录多轮对话消息,每条消息包含角色(role)、内容(content)和工具调用(tool_calls)信息,其中工具调用进一步定义了调用类型(type)和函数详情(包括函数名和参数);tools字段描述可用工具;task字段指明任务类型;num_turns和num_tool_calls分别统计对话轮次和工具调用次数。数据集适用于训练和评估支持工具调用、函数执行或多轮对话的人工智能模型,特别关注工具使用场景下的交互能力。
This dataset is a structured dataset for dialogue and tool calling tasks, containing 1328 training samples. Each sample includes the following core fields: the messages field records multi-turn dialogue messages, each message containing role, content, and tool calls information, where tool calls further define the call type and function details (including function name and parameters); the tools field describes available tools; the task field indicates the task type; num_turns and num_tool_calls respectively count the number of dialogue turns and tool calls. The dataset is suitable for training and evaluating AI models that support tool calls, function execution, or multi-turn dialogues, with a particular focus on interaction capabilities in tool usage scenarios.




