Torch7 && ConvNet
= Torch7 
● Torch7: http://torch.ch/ 
● Easy Install(On Mac OS / Ubuntu): 
https://github.com/torch/ezinstall/ 
● Programming in Lua && Learn Lua in 15 Mins 
http://tylerneylon.com/a/learn-lua/ 
● Machine Learning with Torch7: 
http://code.cogbits.com/wiki/doku.php
= Conv Net – Demo 
● Demo:http://yann.lecun.com/exdb/lenet/ 
● Classification: 
http://www.cs.toronto.edu/~fritz/absps/imagenet 
.pdf
= Conv Net – Intro 
● Large-Scale Visual Recognition With Deep 
Learning: 
http://www.cs.toronto.edu/~ranzato/publication 
s/ranzato_cvpr13.pdf 
● Example:
= Conv Net – Conv
= Conv Net – Pooling
= Conv Net – ReLU && LCN 
● ReLU (Faster training) 
– f(x) = max(0, x) 
● LCN 
– hi+1, x, y= 
hi, x , y−mi, N (x , y) 
σi, N ( x , y )
= Conv Net – Application 
● Joint Trainning of a Convolutional Network and a 
Graphical Model for Human Pose Estimation 
(NIPS 2014) 
● Leaning Human Pose Estimation Features with 
Convolutional Networks (ICLR 2014) 
● More: 
http://cims.nyu.edu/~tompson/cs_portfolio.html
= Conv Net – Application 
● Joint Trainning …... (Section 3.1 , Convolutional 
Network Part-Detector) 
●
= Conv Net – Face Detector 
● Github: 
https://github.com/jonathantompson/geiger_fac 
edetector 
● Others: 
– https://github.com/torch/torch7/wiki/Cheatsheet#de 
mos
= Links 
● http://www.cs.toronto.edu/~ranzato/publication 
s/ranzato_cvpr13.pdf 
● http://deeplearning.stanford.edu/wiki/index.php 
/UFLDL_Tutorial 
● http://cims.nyu.edu/~tompson/cs_portfolio.html 
● http://code.madbits.com/wiki/doku.php

Torch7 and ConvNet

  • 1.
  • 2.
    = Torch7 ●Torch7: http://torch.ch/ ● Easy Install(On Mac OS / Ubuntu): https://github.com/torch/ezinstall/ ● Programming in Lua && Learn Lua in 15 Mins http://tylerneylon.com/a/learn-lua/ ● Machine Learning with Torch7: http://code.cogbits.com/wiki/doku.php
  • 3.
    = Conv Net– Demo ● Demo:http://yann.lecun.com/exdb/lenet/ ● Classification: http://www.cs.toronto.edu/~fritz/absps/imagenet .pdf
  • 4.
    = Conv Net– Intro ● Large-Scale Visual Recognition With Deep Learning: http://www.cs.toronto.edu/~ranzato/publication s/ranzato_cvpr13.pdf ● Example:
  • 5.
    = Conv Net– Conv
  • 6.
    = Conv Net– Pooling
  • 7.
    = Conv Net– ReLU && LCN ● ReLU (Faster training) – f(x) = max(0, x) ● LCN – hi+1, x, y= hi, x , y−mi, N (x , y) σi, N ( x , y )
  • 9.
    = Conv Net– Application ● Joint Trainning of a Convolutional Network and a Graphical Model for Human Pose Estimation (NIPS 2014) ● Leaning Human Pose Estimation Features with Convolutional Networks (ICLR 2014) ● More: http://cims.nyu.edu/~tompson/cs_portfolio.html
  • 10.
    = Conv Net– Application ● Joint Trainning …... (Section 3.1 , Convolutional Network Part-Detector) ●
  • 11.
    = Conv Net– Face Detector ● Github: https://github.com/jonathantompson/geiger_fac edetector ● Others: – https://github.com/torch/torch7/wiki/Cheatsheet#de mos
  • 12.
    = Links ●http://www.cs.toronto.edu/~ranzato/publication s/ranzato_cvpr13.pdf ● http://deeplearning.stanford.edu/wiki/index.php /UFLDL_Tutorial ● http://cims.nyu.edu/~tompson/cs_portfolio.html ● http://code.madbits.com/wiki/doku.php