- 相关
- 目录
- 笔记
- 书签
更多相关文档
-
photographic versus autoradiographic portrayal of dental structure julian d. boyd, m.d
星级： 2 页
-
20. Antigenicity of defatted versus nondefatted pollen extracts James R. Crisp, M.D., and Joseph D. Bullock, M.D., San Francisc
星级： 2 页
-
a.i.
星级： 1 页
-
A.I.来了(下)
星级： 2 页
-
a.i.来了(下)
星级： 3 页
-
a.i.来了(下)
星级： 2 页
更多猜你喜欢
-
fast and flexible probabilistic modelling in python
-
Tacked Weld Point Recognition
-
2018最新版华南理工大学机械与汽车工程学院优秀研究生奖学金评选条例
-
deep learning with spark（基于spark的深度学习）-2017
-
Demonstration of Memory with a Long Short-Term Memory Network in Python - Machine Learning Mastery
-
Recurrent Neural Network models (递归神经网络模型汇总)
- Table of Contents
- About the Author
- About the Technical Reviewer
- Introduction
- Chapter 1: Feedforward Networks
- Chapter 2: Programming Algorithms
- Model Declarations
- Initializations in the Model Constructor
- Finding All Activations
- Activating a Fully Connected Layer
- Activating a Locally Connected Layer
- Activating a Convolutional Layer
- Activating a Pooling Layer
- Evaluating the Criterion
- Evaluating the Gradient
- Gradient for a Fully Connected Layer
- Gradient for a Locally Connected Layer
- Gradient for a Convolutional Layer
- Gradient for a Pooled Layer (Not!)
- Backpropagating Delta from a Nonpooled Layer
- Backpropagating Delta from a Pooled Layer
- Multithreading Gradient Computation
- Chapter 3: CUDA Code
- Weight Layout in the CUDA Implementation
- Global Variables on the Device
- Initialization
- Copying Weights to the Device
- Activating the Output Layer
- Activating Locally Connected and Convolutional Layers
- Activating a Pooled Layer
- SoftMax and Log Likelihood by Reduction
- Computing Delta for the Output Layer
- Backpropagating from a Fully Connected Layer
- Backpropagating from Convolutional and Local Layers
- Backpropagating from a Pooling Layer
- Gradient of a Fully Connected Layer
- Gradient of a Locally Connected or Convolutional Layer
- Fetching the Gradient
- Putting It All Together
- Chapter 4: CONVNET Manual
- Index
暂无笔记
暂无书签
