jacekduszenko/weights-dataset-small

该数据集包含了一系列用于训练机器学习模型的特征，包括训练和测试数据。具体特征包括示例数量、批次数、轮数、批量大小、梯度累积步数、最大训练步数、标签、排名、每个步骤的损失和学习率、LoRa梯度范数、LoRa权重范数以及权重。数据集分为训练集和测试集，分别包含128和32个示例。

The dataset consists of a set of features for training machine learning models, including both training and test data. Specific features include the number of examples, number of batches, number of epochs, batch size, gradient accumulation steps, maximum training steps, label, rank, loss and learning rate at each step, LoRa gradient norm, LoRa weight norm, and weights. The dataset is split into a training set and a test set, containing 128 and 32 examples respectively.