Synaptic Strength For Convolutional Neural Network
摘要 卷积神经网络(Convolutional Neural network, CNNs)既需要大量的计算，又需要大量的内存，这阻碍了它在移动设备上的应用。受到神经科学文献中相关概念的启发，我们提出了突触修剪:一种数据驱动的方法，用一种新提出的参数类突触强度来修剪输入和输出特征图之间的连接。突触强度的设计是基于它传输的信息量来捕捉连接的重要性。实验结果表明了该方法的有效性。在CIFAR-10上，我
相关文章
- 1. Convolutional Neural Network For Sentence Classification
- 2. Convolutional Neural Network Architecture
- 3. Convolutional Neural Network
- 4. CNN(Convolutional Neural Network)
- 5. Convolutional Neural Network - Object Detection
- 6. 【UFLDL】Exercise: Convolutional Neural Network
- 7. Deep Convolutional Neural Network Architecture
- 8. Tensorflow - Implement for a Convolutional Neural Network on MNIST.
- 9. A Cascaded Convolutional Neural Network for Single Image Dehazing
- 10. 《ABCNN: Attention-Based Convolutional Neural Network for Modeling Sentence Pairs》
- 更多相关文章...
- • Swift for 循环 - Swift 教程
- • Scala for循环 - Scala教程
- • Flink 数据传输及反压详解
- • 算法总结-广度优先算法
相关标签/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
欢迎关注本站公众号,获取更多信息
相关文章
- 1. Convolutional Neural Network For Sentence Classification
- 2. Convolutional Neural Network Architecture
- 3. Convolutional Neural Network
- 4. CNN(Convolutional Neural Network)
- 5. Convolutional Neural Network - Object Detection
- 6. 【UFLDL】Exercise: Convolutional Neural Network
- 7. Deep Convolutional Neural Network Architecture
- 8. Tensorflow - Implement for a Convolutional Neural Network on MNIST.
- 9. A Cascaded Convolutional Neural Network for Single Image Dehazing
- 10. 《ABCNN: Attention-Based Convolutional Neural Network for Modeling Sentence Pairs》