All Projects → guanfuchen → cnn_np

guanfuchen / cnn_np

Licence: other
使用numpy构建cnn复习深度学习知识

Programming Languages

Jupyter Notebook
11667 projects
python
139335 projects - #7 most used programming language

Projects that are alternatives of or similar to cnn np

Numpycnn
Building Convolutional Neural Networks From Scratch using NumPy
Stars: ✭ 436 (+1221.21%)
Mutual labels:  numpy, convolution
Dsp Theory
Theory of digital signal processing (DSP): signals, filtration (IIR, FIR, CIC, MAF), transforms (FFT, DFT, Hilbert, Z-transform) etc.
Stars: ✭ 437 (+1224.24%)
Mutual labels:  numpy, convolution
dsp-theory
Theory of digital signal processing (DSP): signals, filtration (IIR, FIR, CIC, MAF), transforms (FFT, DFT, Hilbert, Z-transform) etc.
Stars: ✭ 643 (+1848.48%)
Mutual labels:  numpy, convolution
python-awips
A framework for querying AWIPS meteorological datasets from an EDEX Data Server.
Stars: ✭ 31 (-6.06%)
Mutual labels:  numpy
Dimensionality-reduction-and-classification-on-Hyperspectral-Images-Using-Python
In this repository, You can find the files which implement dimensionality reduction on the hyperspectral image(Indian Pines) with classification.
Stars: ✭ 63 (+90.91%)
Mutual labels:  numpy
ndtest
multi-dimensional statistical test with python
Stars: ✭ 51 (+54.55%)
Mutual labels:  numpy
py-image-search-engine
Python Image Search Engine with OpenCV
Stars: ✭ 37 (+12.12%)
Mutual labels:  numpy
valinvest
A value investing tool based on Warren Buffett, Joseph Piotroski and Benjamin Graham thoughts
Stars: ✭ 84 (+154.55%)
Mutual labels:  numpy
DIP-In-Python
Play around with Pixel in Python
Stars: ✭ 42 (+27.27%)
Mutual labels:  numpy
polytope
Geometric operations on polytopes of any dimension
Stars: ✭ 51 (+54.55%)
Mutual labels:  numpy
dpnp
NumPy drop-in replacement for Intel(R) XPUs
Stars: ✭ 42 (+27.27%)
Mutual labels:  numpy
spyndex
Awesome Spectral Indices in Python.
Stars: ✭ 56 (+69.7%)
Mutual labels:  numpy
ai-virtual-mouse
Developed an AI-based system to control the mouse cursor using Python and OpenCV with the real-time camera. Fingertip location is mapped to RGB images to control the mouse cursor.
Stars: ✭ 59 (+78.79%)
Mutual labels:  numpy
MachineLearning
机器学习教程,本教程包含基于numpy、sklearn与tensorflow机器学习,也会包含利用spark、flink加快模型训练等用法。本着能够较全的引导读者入门机器学习。
Stars: ✭ 23 (-30.3%)
Mutual labels:  numpy
slam-python
SLAM - Simultaneous localization and mapping using OpenCV and NumPy.
Stars: ✭ 80 (+142.42%)
Mutual labels:  numpy
SciCompforChemists
Scientific Computing for Chemists text for teaching basic computing skills to chemistry students using Python, Jupyter notebooks, and the SciPy stack. This text makes use of a variety of packages including NumPy, SciPy, matplotlib, pandas, seaborn, NMRglue, SymPy, scikit-image, and scikit-learn.
Stars: ✭ 65 (+96.97%)
Mutual labels:  numpy
pnumpy
Parallel NumPy seamlessly speeds up NumPy for large arrays (64K+ elements) with no change required to existing code.
Stars: ✭ 45 (+36.36%)
Mutual labels:  numpy
CS231n
PyTorch/Tensorflow solutions for Stanford's CS231n: "CNNs for Visual Recognition"
Stars: ✭ 47 (+42.42%)
Mutual labels:  numpy
Deep-Surveillance-Monitor-Facial-Emotion-Age-Gender-Recognition-System
Computer Vision module for detecting emotion, age and gender of a person in any given image, video or real time webcam. A custom VGG16 model was developed and trained on open source facial datasets downloaded from Kaggle and IMDB. OpenCV,dlib & keras were used to aid facial detection and video processing. The final system can detect the emotion,…
Stars: ✭ 60 (+81.82%)
Mutual labels:  numpy
vinum
Vinum is a SQL processor for Python, designed for data analysis workflows and in-memory analytics.
Stars: ✭ 57 (+72.73%)
Mutual labels:  numpy

cnn_np

本仓库主要使用numpy来实现深度学习中常用的网络结构,复习深度学习相关知识。

使用构造的网络模块搭建LeNet-5(结构参考Caffe实现)训练手写数字识别。


卷积

卷积模块实现时,使用标准模块测试输入输出。

卷积前向传播较为简单,只需要将权值转换为列向量,图像转换为列向量,然后按照矩阵乘法同时加上偏置在转换为图像即可。

卷积反向传播较难,首先通过误差计算权值和偏置的梯度,然后计算该误差下一传播过程中的误差,该计算中注意误差需要通过pad然后转换为列向量,和权重上下翻转后相乘即可。

pytorch卷积实现conv.py

在 Caffe 中如何计算卷积? 贾扬清讲解caffe中如何计算卷积。

Convolutional Neural Networks backpropagation: from intuition to derivation


池化层

本仓库实现了MaxPool2d。

Convnet: Implementing Maxpool Layer with Numpy


全连接层

全连接层

linear.py


Softmax层

Softmax层当前神经元的输出对当前的神经元的导数之和$-a_ia_j$以及$a_i(1-a_i)$。

softmax函数与交叉熵的反向梯度传导 本文介绍了softmax层的反向传播。


激活函数层

activation.py pytorch中激活函数设计参考。


损失函数层

CrossEntropy层使用了one-hot编码的target。

pytorch loss function 总结 其中介绍了pytorch中各个损失函数的关系,本仓库实现CrossEntropy包含LogSoftmax。

loss.py pytorch中相应的loss设计参考。

One Hot Encoding using numpy

numpy之one-hot


实现日志

  • Conv2d
  • Pool2d
  • Linear
  • Softmax
  • CrossEntropy
  • ReLU
  • LReLU

参考链接

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].