Difference between revisions of "Getting started with Deep Learning"

From DeepSense Docs
Jump to: navigation, search
(→‎Note)
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
<div class="noautonum">
 
<div class="noautonum">
  
−
== Note ==
 
−
On June 26 we updated the GPU compute nodes to a new version of IBM Watson Machine Learning Accelerator. This changes the way you access deep learning packages like Tensorflow and Pytorch. Instead of "activating" these packages, you will be able to install new versions directly in your anaconda environment.
 
−
 
−
We are actively updating the wiki documentation to explain the new method of accessing deep learning packages. Please bear with us during these updates as some documentation may still refer to the old method of "activating" deep learning packages
 
  
 
== 1. Get started with DeepSense ==
 
== 1. Get started with DeepSense ==
  
−
Follow all the steps from [[Getting started]]. This tutorial assumes you can log on to the DeepSense compute platform and have a version of Anaconda python on your path. We recommend installing Anaconda in your home directory before starting this tutorial (See [[Installing local software]]).
+
Follow all the steps from [[Getting started]]. This tutorial assumes you can log on to the DeepSense compute platform and have a version of Anaconda python on your path. We recommend installing Anaconda in your home directory before starting this tutorial (See [[Installing Software]]).
  
 
== 2. Prepare Caffe and download Caffe samples to your home directory ==
 
== 2. Prepare Caffe and download Caffe samples to your home directory ==
Line 14: Line 10:
 
(New method)
 
(New method)
  
−
Activate your anaconda environment. See [[Installing local software]] for how to create an environment. We will assume you have created one called "caffe".
+
Activate your anaconda environment. See [[Installing Software]] for how to create an environment. We will assume you have created one called "caffe".
 
  conda activate caffe
 
  conda activate caffe
 
Add the IBM-AI anaconda channel if you have not done so already
 
Add the IBM-AI anaconda channel if you have not done so already
Line 27: Line 23:
 
<code>/opt/DL/caffe/bin/caffe-install-samples</code>
 
<code>/opt/DL/caffe/bin/caffe-install-samples</code>
  
−
== 3. Request an interactive session on a GPU compute node ==
 
−
 
−
<!-- TODO: still need to set up queues to fairly share GPUs -->
 
−
<!-- TODO: write instructions doing this with regular LSF without an interactive session. We don't want to encourage everyone to use interactive sessions -->
 
−
<code>bsub -Is -gpu - bash</code>
 
−
 
−
== 4. Start a python2 Jupyter notebook ==
 
−
 
−
=== (Old method only) Source the Caffe deep learning toolkit ===
 
−
 
−
<code>source /opt/DL/caffe/bin/caffe-activate</code>
 
−
 
−
=== Start the notebook ===
 
  
−
<code>jupyter notebook --no-browser --ip=0.0.0.0</code>
+
== 3. More information ==
−
 
 
−
=== Sample output ===
 
−
<pre>[I 13:32:23.937 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
 
−
[C 13:32:23.937 NotebookApp]
 
−
   
 
−
    Copy/paste this URL into your browser when you connect for the first time,
 
−
    to login with a token:
 
−
        http://ds-cmgpu-04:8888/?token=68042f40a10b500f3747ae0a232ee209fa4bf1aa384d29ba&token=68042f40a10b500f3747ae0a232ee209fa4bf1aa384d29ba
 
−
</pre>
 
−
 
 
−
=== Copy the URL, host, and port ===
 
−
 
 
−
Copy the URL but don’t paste it in your browser yet.
 
−
 
 
−
Make a note of which compute host and port the notebook is running on (e.g. host ds-cmgpu-04 and port 8888 in this case)
 
−
 
 
−
== 5. Port Forwarding ==
 
−
 
 
−
In a separate terminal window from your local computer, forward your local port to the remote host.
 
−
 
 
−
=== ssh command port forwarding ===
 
−
 
 
−
<code> ssh -l <username> login1.deepsense.ca -L <local_port>:<remote_host>:<remote_port></code>
 
−
 
 
−
for example, <code>ssh -l user1 login1.deepsense.ca -L 8888:ds-cmgpu-04:8888</code>
 
−
 
 
−
Note that you may need to use a different <local_port> than 8888 if you have other web services running on your local computer. In particular, if you run a jupyter notebook locally then it will use port 8888 and you will try to connect to the local jupyter notebook instead of the cluster notebook. In this case close your port forwarding and try again with 8889 or another unused port.
 
−
 
 
−
=== PuTTY port forwarding on Windows ===
 
−
 
 
−
If you are using a PuTTY terminal from a Windows computer to access DeepSense then you can still forward ports.
 
−
 
 
−
Before starting your session, scroll down to the option <code>Connection->SSH->Tunnels</code> in the Category pane.
 
−
 
 
−
Enter the <code>local_port</code> in the <code>Source port</code> field. For example, <code>8888</code>.
 
−
 
 
−
Enter <code><remote_host>:<remote_port></code> in the Destination field. For example, <code>ds-cmgpu-04:8888</code>.
 
−
 
 
−
Press the <code>Add</code> button to add the port forwarding rule to your PuTTY session.
 
−
 
 
−
Finally, open the session as usual.
 
−
 
 
−
== 6. Open the desired sample notebook ==
 
−
 
 
−
Enter the copied URL in your web browser but change the remote host name to “localhost” before pressing enter.
 
−
 
 
−
e.g <code>http://localhost:8888/?token=68042f40a10b500f3747ae0a232ee209fa4bf1aa384d29ba&token=68042f40a10b500f3747ae0a232ee209fa4bf1aa384d29ba</code>
 
−
 
 
−
'''Note''': On our macs, this worked in Chrome, but not in Safari.  Unfortunately, there was no error reported, it simply could not connect.
 
−
 
 
−
 
 
−
Be sure to enter the location of the “caffe-samples” directory in your home directory as your caffe-root in the Caffe example notebooks.
 
−
 
 
−
== 7. Enjoy Deep Learning on DeepSense! ==
 
−
 
 
−
== 8. More information ==
 
  
 
Go to Caffe's [http://caffe.berkeleyvision.org/ website] for tutorials and example programs that you can run to get started.
 
Go to Caffe's [http://caffe.berkeleyvision.org/ website] for tutorials and example programs that you can run to get started.
Line 106: Line 33:
 
[http://caffe.berkeleyvision.org/gathered/examples/cifar10.html CIFAR-10 tutorial] - Train a convolutional neural network to classify small images.
 
[http://caffe.berkeleyvision.org/gathered/examples/cifar10.html CIFAR-10 tutorial] - Train a convolutional neural network to classify small images.
  
−
== 9. Using another deep learning toolkit such as Tensorflow ==
+
== 4. Using another deep learning toolkit such as Tensorflow ==
  
  
Line 130: Line 57:
  
 
   https://github.com/aymericdamien/TensorFlow-Examples
 
   https://github.com/aymericdamien/TensorFlow-Examples
−
 
−
 
−
 
−
 
−
 
−
 
  
 
</div> <!-- autonum -->
 
</div> <!-- autonum -->

Latest revision as of 14:26, 3 December 2020


1. Get started with DeepSense

Follow all the steps from Getting started. This tutorial assumes you can log on to the DeepSense compute platform and have a version of Anaconda python on your path. We recommend installing Anaconda in your home directory before starting this tutorial (See Installing Software).

2. Prepare Caffe and download Caffe samples to your home directory

(New method)

Activate your anaconda environment. See Installing Software for how to create an environment. We will assume you have created one called "caffe". 

conda activate caffe

Add the IBM-AI anaconda channel if you have not done so already 

conda config --prepend channels https://public.dhe.ibm.com/ibmdl/export/pub/software/server/ibm-ai/conda/

Install caffe if you have not done so already 

conda install caffe

Install the caffe samples 

caffe-install-samples

(Old method)

/opt/DL/caffe/bin/caffe-install-samples


3. More information

Go to Caffe's website for tutorials and example programs that you can run to get started. See the following links to a couple of the example programs:

LeNet MNIST Tutorial - Train a neural network to understand handwritten digits.

CIFAR-10 tutorial - Train a convolutional neural network to classify small images.

4. Using another deep learning toolkit such as Tensorflow

(New Method)

(Old Method)

  • Ensure any Anaconda dependencies are installed
    • for tensorflow, run /opt/DL/tensorflow/bin/install_dependencies
  • Source the appropriate toolkit instead of caffe-activate
    • e.g. source /opt/DL/tensorflow/bin/tensorflow-activate
  • Download example notebooks for the deep learning toolkit to your home directory,

The TensorFlow home page has various information, including Tutorials, How-To documents, and a Getting Started guide.

Additional tutorials and examples are available from the community, for example: 

 https://github.com/nlintz/TensorFlow-Tutorials

 https://github.com/aymericdamien/TensorFlow-Examples

Retrieved from "https://docs.deepsense.ca/index.php?title=Getting_started_with_Deep_Learning&oldid=327"