stefanocarrera/autophagycode_D_he_train-mercury_Qwen3-4B_strategy_trust_t1.1_g8_run0
收藏资源简介:
该数据集是一个用于任务导向的代码生成或自然语言处理的数据集,包含164个训练示例。每个示例具有以下特征:task_id(唯一任务标识符)、entry_point(任务入口点)、prompt(输入提示文本)、completion(对应的完成文本)、top_k_progression(top-k进展信息)和test(测试相关数据),所有特征均为字符串类型。数据集仅提供训练分割,总大小约为4.36 MB,下载大小约为1.04 MB,数据文件位于data/train-*路径下。该数据集可能用于评估或训练模型在代码生成、任务完成等场景中的性能。
This dataset is designed for task-oriented code generation or natural language processing, containing 164 training examples. Each example includes the following features: task_id (unique task identifier), entry_point (task entry point), prompt (input prompt text), completion (corresponding completion text), top_k_progression (top-k progression information), and test (test-related data), all with string data types. The dataset provides only a training split, with a total size of approximately 4.36 MB and a download size of approximately 1.04 MB. Data files are located at the path data/train-*. It may be used for evaluating or training models in scenarios such as code generation and task completion.




