Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update th_tutorials/0_getstarted/README #18

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Update th_tutorials/0_getstarted/README #18

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 5 additions & 5 deletions th_tutorials/0_getstarted/README.html
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,7 +32,7 @@ <h1 id="getting-started-with-torch">Getting Started with Torch</h1>
<p>Torch7 is developed at <a href="http://www.idiap.ch/">Idiap Research Institute</a>, <a href="http://www.cs.nyu.edu/~yann/">New York University</a> and <a href="http://www.nec-labs.com/">NEC Laboratories America</a>.</p>
<p>At this time, Torch7 is maintained by <a href="http://ronan.collobert.com/">Ronan Collobert</a>, <a href="http://www.clement.farabet.net/">Clement Farabet</a> and <a href="http://koray.kavukcuoglu.org/">Koray Kavukcuoglu</a>. We also use the Qt &lt;-&gt; Lua interface from <a href="http://leon.bottou.org/">Leon Bottou</a>.</p>
<h2 id="installing-torch">Installing Torch</h2>
<p>For these tutorials, we assume that Torch is already installed, as well as extra packages (<em>image</em>, <em>nnx</em>, <em>unsup</em>). If you want to install Torch on your machine (which I recommend for the next sessions), follow the instructions available <a href="http://www.torch.ch/manual/install/index">here</a>. A more condensed version can also be found <a href="http://code.cogbits.com/">here</a>.</p>
<p>For these tutorials, we assume that Torch is already installed, as well as extra packages (<em>image</em>, <em>nnx</em>, <em>unsup</em>). If you want to install Torch on your machine (which I recommend for the next sessions), follow the instructions available <a href="http://torch.ch/docs/getting-started.html">here</a>. A more condensed version can also be found <a href="http://code.cogbits.com/">here</a>.</p>
<h2 id="virtual-machine-with-torch7-pre-installed">Virtual Machine with Torch7 pre-installed</h2>
<p>I'm distributing a virtual machine, with Ubuntu 12, and Torch7 pre-installed, <a href="http://data.neuflow.org/share/Torch7.ova">here</a>. You can import this virtual machine using <a href="https://www.virtualbox.org/">VirtualBox</a>. That's probably the best way to get started with Torch7.</p>
<h2 id="ec2-machine-pre-built-torch-for-the-tutorials">EC2 Machine: Pre-built Torch for the tutorials</h2>
Expand Down Expand Up @@ -158,11 +158,11 @@ <h2 id="getting-help">Getting help</h2>
</pre></td><td class="sourceCode"><pre><code class="sourceCode lua"><span class="kw">t7</span><span class="ot">&gt;</span> help<span class="ot">(</span><span class="kw">torch</span><span class="ot">.</span><span class="kw">randn</span><span class="ot">)</span>

<span class="er">++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++</span><span class="ot">+</span>
<span class="ot">[</span><span class="kw">res</span><span class="ot">]</span> <span class="kw">torch</span><span class="ot">.</span>randn<span class="ot">(</span> <span class="ot">[</span><span class="kw">res</span>,<span class="ot">]</span> <span class="kw">m</span> <span class="ot">[</span>, <span class="kw">n</span>, <span class="kw">k</span>, <span class="ot">...])</span>
<span class="kw">y</span><span class="ot">=</span><span class="kw">torch</span><span class="ot">.</span>randn<span class="ot">(</span><span class="kw">n</span><span class="ot">)</span> <span class="kw">returns</span> <span class="kw">a</span> <span class="kw">one</span><span class="ot">-</span><span class="kw">dimensional</span> <span class="kw">tensor</span> <span class="kw">of</span> <span class="kw">size</span> <span class="kw">n</span> <span class="kw">filled</span>
<span class="kw">with</span> <span class="fu">random</span> <span class="kw">numbers</span> <span class="kw">from</span> <span class="kw">a</span> <span class="kw">normal</span> <span class="kw">distribution</span> <span class="kw">with</span> <span class="kw">mean</span> <span class="kw">zero</span> <span class="kw">and</span> <span class="kw">variance</span>
<span class="ot">[</span><span class="kw">res</span><span class="ot">]</span> <span class="kw">torch</span><span class="ot">.</span>randn<span class="ot">(</span> <span class="ot">[</span><span class="kw">res</span>,<span class="ot">]</span> <span class="kw">m</span> <span class="ot">[</span>, <span class="kw">n</span>, <span class="kw">k</span>, <span class="ot">...])</span>
<span class="kw">y</span><span class="ot">=</span><span class="kw">torch</span><span class="ot">.</span>randn<span class="ot">(</span><span class="kw">n</span><span class="ot">)</span> <span class="kw">returns</span> <span class="kw">a</span> <span class="kw">one</span><span class="ot">-</span><span class="kw">dimensional</span> <span class="kw">tensor</span> <span class="kw">of</span> <span class="kw">size</span> <span class="kw">n</span> <span class="kw">filled</span>
<span class="kw">with</span> <span class="fu">random</span> <span class="kw">numbers</span> <span class="kw">from</span> <span class="kw">a</span> <span class="kw">normal</span> <span class="kw">distribution</span> <span class="kw">with</span> <span class="kw">mean</span> <span class="kw">zero</span> <span class="kw">and</span> <span class="kw">variance</span>
<span class="kw">one</span><span class="ot">.</span>
<span class="kw">y</span><span class="ot">=</span><span class="kw">torch</span><span class="ot">.</span>randn<span class="ot">(</span><span class="kw">m</span>,<span class="kw">n</span><span class="ot">)</span> <span class="kw">returns</span> <span class="kw">a</span> <span class="kw">mxn</span> <span class="kw">tensor</span> <span class="kw">of</span> <span class="fu">random</span> <span class="kw">numbers</span> <span class="kw">from</span> <span class="kw">a</span> <span class="kw">normal</span>
<span class="kw">y</span><span class="ot">=</span><span class="kw">torch</span><span class="ot">.</span>randn<span class="ot">(</span><span class="kw">m</span>,<span class="kw">n</span><span class="ot">)</span> <span class="kw">returns</span> <span class="kw">a</span> <span class="kw">mxn</span> <span class="kw">tensor</span> <span class="kw">of</span> <span class="fu">random</span> <span class="kw">numbers</span> <span class="kw">from</span> <span class="kw">a</span> <span class="kw">normal</span>
<span class="kw">distribution</span> <span class="kw">with</span> <span class="kw">mean</span> <span class="kw">zero</span> <span class="kw">and</span> <span class="kw">variance</span> <span class="kw">one</span><span class="ot">.</span>
<span class="er">++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++</span><span class="ot">+</span></code></pre></td></tr></table>
<h2 id="going-further">Going Further</h2>
Expand Down
46 changes: 23 additions & 23 deletions th_tutorials/0_getstarted/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,21 +2,21 @@ Getting Started with Torch
==========================

Torch7 provides a Matlab-like environment for state-of-the-art machine
learning algorithms. It is easy to use and provides a very efficient
learning algorithms. It is easy to use and provides a very efficient
implementation, thanks to an easy and fast scripting language,
[Lua](http://www.lua.org/) and an underlying C/OpenMP/CUDA implementation.

Torch7 is developed at
[Idiap Research Institute](http://www.idiap.ch/),
Torch7 is developed at
[Idiap Research Institute](http://www.idiap.ch/),
[New York University](http://www.cs.nyu.edu/~yann/) and
[NEC Laboratories America](http://www.nec-labs.com/).
[NEC Laboratories America](http://www.nec-labs.com/).

At this time, Torch7 is maintained by
[Ronan Collobert](http://ronan.collobert.com/),
At this time, Torch7 is maintained by
[Ronan Collobert](http://ronan.collobert.com/),
[Clement Farabet](http://www.clement.farabet.net/)
and
and
[Koray Kavukcuoglu](http://koray.kavukcuoglu.org/).
We also use the Qt <-> Lua interface from
We also use the Qt <-> Lua interface from
[Leon Bottou](http://leon.bottou.org/).

Installing Torch
Expand All @@ -25,7 +25,7 @@ Installing Torch
For these tutorials, we assume that Torch is already installed, as well as extra
packages (_image_, _nnx_, _unsup_). If you want to install Torch on your machine
(which I recommend for the next sessions), follow the instructions available
[here](http://www.torch.ch/manual/install/index). A more condensed version can
[here](http://torch.ch/docs/getting-started.html). A more condensed version can
also be found [here](http://code.cogbits.com/).

Virtual Machine with Torch7 pre-installed
Expand All @@ -39,13 +39,13 @@ the best way to get started with Torch7.
EC2 Machine: Pre-built Torch for the tutorials
----------------------------------------------

For now, we have set up a server (Amazon EC2), which contains a pre-built version
For now, we have set up a server (Amazon EC2), which contains a pre-built version
of Torch, and all the packages necessary to run the demos/tutorials provided.
The IP number of the server will change periodically, so we'll just tell you
what it is when you need it.

We will provide you with an identity file `ipam_identity.pem` file, which you'll use
to log into the EC2 machine. Once you've got the `ipam_identity.pem` file and the
to log into the EC2 machine. Once you've got the `ipam_identity.pem` file and the
`ADDRESS` of the server, pick a unique `USERNAME`, for your use within the "student"
account, and if you have a linux or OSX computer type:

Expand Down Expand Up @@ -88,7 +88,7 @@ t7> print 'something'
something
t7> = 'something'
something
t7>
t7>
```

By default, the interpreter preloads the torch and plotting packages. Extra
Expand Down Expand Up @@ -125,7 +125,7 @@ t7> image.display{image=n.output, padding=2, zoom=0.25, legend='states'}

![](https://github.com/clementfarabet/ipam-tutorials/raw/master/th_tutorials/0_getstarted/img/states.png)

Note: most functions in Torch support named arguments. In the example above,
Note: most functions in Torch support named arguments. In the example above,
image.display can be called with arguments in order, or named. When using named
arguments, notice that we replace the parenthesis by curly brackets. Curly
brackets are used to describe Lua's most general data type: the table.
Expand Down Expand Up @@ -158,9 +158,9 @@ t7> dofile 'getstarted.lua'
Getting help
------------

Torch's documentation is work in progress, but you can already get help for most function
Torch's documentation is work in progress, but you can already get help for most function
provided in the official packages (_torch_, _nn_, _gnuplot_, _image_). Documentation
is installed on your machine, along with Torch's packages. The default location is
is installed on your machine, along with Torch's packages. The default location is
[/usr/local/share/torch/html/index.html](file:///usr/local/share/torch/html/index.html). It
is also mirrored online [here](http://www.torch.ch/manual), and can be triggered from
your interpreter by doing:
Expand Down Expand Up @@ -188,10 +188,10 @@ t7> help(torch.randn)

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[res] torch.randn( [res,] m [, n, k, ...])
y=torch.randn(n) returns a one-dimensional tensor of size n filled
with random numbers from a normal distribution with mean zero and variance
y=torch.randn(n) returns a one-dimensional tensor of size n filled
with random numbers from a normal distribution with mean zero and variance
one.
y=torch.randn(m,n) returns a mxn tensor of random numbers from a normal
y=torch.randn(m,n) returns a mxn tensor of random numbers from a normal
distribution with mean zero and variance one.
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
```
Expand All @@ -201,7 +201,7 @@ Going Further

Now that you have all the basic pieces, I invite you to take a look at Torch's
[basic tutorial/introduction](http://www.torch.ch/manual/tutorial/index). In
particular, take a look at:
particular, take a look at:

* a very quick [Lua primer](http://www.torch.ch/manual/tutorial/index#lua_basics),
* a fast intro to [Torch types](http://www.torch.ch/manual/tutorial/index#torch_basicsplaying_with_tensors),
Expand All @@ -213,14 +213,14 @@ provide code for all sorts of things, including image processing, video processi
graphical models (CRFs, ...), parallel computing, camera interfaces, ... A complete
list of packages we support can be found [here](http://code.cogbits.com/packages/).
Torch has a built-in package management system that makes it very easy for anyone
to get and develop new packages, which can be shared easily, using, for example
to get and develop new packages, which can be shared easily, using, for example
[GitHub](https://github.com/) as a distribution platform. More details about this
[here](http://www.torch.ch/manual/install/index#the_torch_package_management_system).

(Easy) Exercise
---------------

Ok it's day 1, we just got started, but if we have enough time, let's try to anticipate
Ok it's day 1, we just got started, but if we have enough time, let's try to anticipate
what we'll have to do in the next days, with a simple exercise.

In any machine learning task, data plays a central role. The most basic thing to
Expand All @@ -243,7 +243,7 @@ t7> = train
{[data] = ByteTensor - size: 60000x1x32x32
[labels] = ByteTensor - size: 60000}
```

Data needs to be normalized. Verify the initial data range, and
write the necessary code to insure that the data has zero mean and
unit norm. Depending on the task, you might want to normalize the
Expand All @@ -255,5 +255,5 @@ efficient numeric routines (ala Matlab). If you can't figure out how
to do it, check out this script: `../1_supervised/A_slicing.lua`.

Last little challenge, try to display a subset of the training
images. By now you should have seen enough stuff to be able to
images. By now you should have seen enough stuff to be able to
do that.