【Fast Algorithms for Convolutional Neural Networks】论文笔记
这篇文章基于Winograd开创的最小滤波算法，介绍了一类新的卷积神经网络快速算法。 针对卷积计算方式而言，与直接卷积相比，该算法可以将卷积层的算术复杂度降低4倍。 算法部分 对于最小滤波算法而言，使用r维滤波器计算m个输出(称之为F(m,r))需要m+r-1次乘法计算。 标准计算而言的话，需要m×r次乘法计算。 对于1-D卷积： 输入x=[1,2,3,4]，卷积核大小为3×1(r=3)，k=[1
相关文章
- 1. 论文阅读-Fast Algorithms for Convolutional Neural Networks
- 2. Learning Convolutional Neural Networks for Graphs论文笔记
- 3. [论文笔记] [2016] [NIPS] Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
- 4. 论文笔记：Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
- 5. Diffusion-Convolutional Neural Networks论文笔记
- 6. 论文笔记——Factorized Convolutional Neural Networks
- 7. 论文笔记：Recurrent Convolutional Neural Networks for Text Classification
- 8. 论文《Convolutional Neural Networks for Sentence Classification》总结
- 9. 论文阅读：《Convolutional Neural Networks for Sentence Classification》
- 10. 《Convolutional Neural Networks for Sentence Classification》论文研读
- 更多相关文章...
- • ASP.NET Razor - 标记 - ASP.NET 教程
- • Scala for循环 - Scala教程
- • Tomcat学习笔记（史上最全tomcat学习笔记）
- • Scala 中文乱码解决
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. 论文阅读-Fast Algorithms for Convolutional Neural Networks
- 2. Learning Convolutional Neural Networks for Graphs论文笔记
- 3. [论文笔记] [2016] [NIPS] Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
- 4. 论文笔记：Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering
- 5. Diffusion-Convolutional Neural Networks论文笔记
- 6. 论文笔记——Factorized Convolutional Neural Networks
- 7. 论文笔记：Recurrent Convolutional Neural Networks for Text Classification
- 8. 论文《Convolutional Neural Networks for Sentence Classification》总结
- 9. 论文阅读：《Convolutional Neural Networks for Sentence Classification》
- 10. 《Convolutional Neural Networks for Sentence Classification》论文研读