mazkooleg/digit_mask_false_positive_cv12_ft_ensemble_distilled_mfcc

该数据集包含三个主要特征：标签（label）、逻辑值（logits）和梅尔频率倒谱系数（mfcc）。数据集被分为训练集、测试集和验证集，分别包含1,825,800、6,086和5,276个样本。训练集的大小为9,297,201,825字节，测试集为30,990,673字节，验证集为26,866,052字节。数据集的下载大小为9,557,109,006字节，总大小为9,355,058,550字节。数据文件路径分别为：训练集（data/train-*）、测试集（data/test-*）和验证集（data/validation-*）。

The dataset includes three main features: label, logits, and Mel-frequency cepstral coefficients (mfcc). The dataset is divided into training, test, and validation sets, containing 1,825,800, 6,086, and 5,276 samples respectively. The training set size is 9,297,201,825 bytes, the test set is 30,990,673 bytes, and the validation set is 26,866,052 bytes. The download size of the dataset is 9,557,109,006 bytes, and the total size is 9,355,058,550 bytes. The data file paths are: training set (data/train-*), test set (data/test-*), and validation set (data/validation-*).