Speech Commands
收藏arXiv2025-09-30 收录
下载链接：
https://www.tensorflow.org/lite/models/modify/model_maker#dataset
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了51,088个特定关键词的语音样本，用于训练卷积神经网络（CNN）。此外，数据集还包括了30个不同的关键词，并提取了梅尔频谱图以供CNN训练使用。在经过10,000次训练迭代后，该模型的准确率达到了89.12%。该任务的目的是实现语音识别。
This dataset contains 51,088 speech samples associated with 30 specific keywords, intended for training Convolutional Neural Networks (CNNs). Additionally, the dataset includes these 30 distinct keywords, and their Mel spectrograms have been extracted for CNN training. After 10,000 training iterations, the trained model achieved an accuracy of 89.12%. The goal of this task is speech recognition.
提供机构：
Google
