yuvals2/crowd-transcribe-v5-processed

该数据集是一个包含希伯来语（he）数据的数据集，主要用于处理序列数据。数据集的特征包括input_features（浮点数序列）、labels（整数序列）和input_length（浮点数）。数据集分为训练集和测试集，训练集包含203,826个样本，测试集包含20,709个样本。总数据集大小估计为353,299,638,170字节。

This dataset contains Hebrew (he) language data and is primarily used for sequence data processing. The dataset features include input_features (sequence of float32), labels (sequence of int64), and input_length (float64). The dataset is divided into a training set with 203,826 examples and a test set with 20,709 examples. The total dataset size is estimated to be 353,299,638,170 bytes.