ngtranai09/HybridSmoothLabels

该数据集包含模型训练和评估所需的各种参数和指标，如模型名称、训练集和测试集大小、自动查找批次大小、梯度累积步数、学习率、日志记录步数、学习率调度器类型、训练周期数、优化器类型、输出目录、报告目标、保存策略、保存总限制、随机种子、预热步数、权重衰减等。此外，还包括序列参数、参数数量、可训练参数数量、内存分配情况、训练时间以及性能指标（包括准确率、宏平均F1分数、加权平均F1分数、精确率、召回率）。数据集分为训练集，其大小为10954字节，包含37个示例。

The dataset contains various parameters and metrics necessary for model training and evaluation, such as model name, training and test set sizes, auto-find batch size, gradient accumulation steps, learning rate, logging steps, learning rate scheduler type, number of training epochs, optimizer type, output directory, reporting target, save strategy, total save limit, random seed, warmup steps, weight decay, etc. It also includes sequence parameters, number of parameters, number of trainable parameters, memory allocation, training time, and performance metrics (including accuracy, macro F1 score, weighted average F1 score, precision, recall). The dataset is split into a training set, which is 10954 bytes in size and contains 37 examples.