Condense-AI/benchmark-condense-v0.1

该数据集是通过子网中的挑战生成器生成的，主要用于测试和基准测试。数据集包含任务、ID、上下文、激活提示、预期完成、模型ID和最大字符数等特征。数据集分为训练集，包含150,000个示例，总大小为2,328,171,903字节。生成数据集的脚本和基准测试的输出结果也包含在README文件中。The Synthetic Challenge - Neural Condense Subnet - Bittensor dataset is a challenge dataset generated in the subnet. The dataset includes features such as task type, ID, context, activation prompt, expected completion, model ID, and maximum characters. The dataset is divided into a training set with 150,000 samples. The dataset can be reproduced by running the provided Python script, which also generates a file named benchmark_dataset.json.