Author: wangwei
Date: Thu Oct 6 03:20:51 2016
New Revision: 1763496
URL: http://svn.apache.org/viewvc?rev=1763496&view=rev
Log:
update the dev schedule and add a FAQ entry into installation page
Modified:
incubator/singa/site/trunk/en/develop/schedule.html
incubator/singa/site/trunk/en/docs/installation.html
Modified: incubator/singa/site/trunk/en/develop/schedule.html
URL:
http://svn.apache.org/viewvc/incubator/singa/site/trunk/en/develop/schedule.html?rev=1763496&r1=1763495&r2=1763496&view=diff
==============================================================================
--- incubator/singa/site/trunk/en/develop/schedule.html (original)
+++ incubator/singa/site/trunk/en/develop/schedule.html Thu Oct 6 03:20:51 2016
@@ -150,291 +150,172 @@
<h1>Development Schedule<a class="headerlink" href="#development-schedule"
title="Permalink to this headline">¶</a></h1>
<table border="1" class="docutils">
<colgroup>
-<col width="25%" />
-<col width="25%" />
-<col width="25%" />
-<col width="25%" />
+<col width="33%" />
+<col width="33%" />
+<col width="33%" />
</colgroup>
<thead valign="bottom">
<tr class="row-odd"><th class="head">Release</th>
<th class="head">Module</th>
<th class="head">Feature</th>
-<th class="head">Status</th>
</tr>
</thead>
<tbody valign="top">
-<tr class="row-even"><td><blockquote class="first last">
-<div>0.1 Sep 2015</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Neural Network</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Feed forward neural network, including CNN, MLP</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<tr class="row-even"><td>0.1 Sep 2015</td>
+<td>Neural Network</td>
+<td>Feed forward neural network, including CNN, MLP</td>
</tr>
<tr class="row-odd"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>RBM-like model, including RBM</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<td>RBM-like model, including RBM</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Recurrent neural network, including standard RNN</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<td>Recurrent neural network, including standard RNN</td>
</tr>
<tr class="row-odd"><td> </td>
-<td><blockquote class="first last">
-<div>Architecture</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>One worker group on single node (with data partition)</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<td>Architecture</td>
+<td>One worker group on single node (with data partition)</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Multi worker groups on single node using [Hogwild](<a class="reference
external"
href="http://www.eecs.berkeley.edu/~brecht/papers/hogwildTR.pdf";>http://www.eecs.berkeley.edu/~brecht/papers/hogwildTR.pdf</a>)</div></blockquote>
-</td>
-<td>done</td>
+<td>Multi worker groups on single node using <a class="reference external"
href="http://www.eecs.berkeley.edu/~brecht/papers/hogwildTR.pdf";>Hogwild</a></td>
</tr>
<tr class="row-odd"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Distributed Hogwild</div></blockquote>
-</td>
-<td>done</td>
+<td>Distributed Hogwild</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Multi groups across nodes, like [Downpour](<a class="reference external"
href="http://papers.nips.cc/paper/4687-large-scale-distributed-deep-networks";>http://papers.nips.cc/paper/4687-large-scale-distributed-deep-networks</a>)</div></blockquote>
-</td>
-<td>done</td>
+<td>Multi groups across nodes, like <a class="reference external"
href="http://papers.nips.cc/paper/4687-large-scale-distributed-deep-networks";>Downpour</a></td>
</tr>
<tr class="row-odd"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>All-Reduce training architecture like [DeepImage](<a class="reference
external"
href="http://arxiv.org/abs/1501.02876";>http://arxiv.org/abs/1501.02876</a>)</div></blockquote>
-</td>
-<td>done</td>
+<td>All-Reduce training architecture like <a class="reference external"
href="http://arxiv.org/abs/1501.02876";>DeepImage</a></td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Load-balance among servers</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<td>Load-balance among servers</td>
</tr>
<tr class="row-odd"><td> </td>
-<td><blockquote class="first last">
-<div>Failure recovery</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Checkpoint and restore</div></blockquote>
-</td>
-<td>done</td>
+<td>Failure recovery</td>
+<td>Checkpoint and restore</td>
</tr>
<tr class="row-even"><td> </td>
-<td><blockquote class="first last">
-<div>Tools</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Installation with GNU auto tools</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<td>Tools</td>
+<td>Installation with GNU auto Tools</td>
</tr>
<tr class="row-odd"><td>0.2 Jan 2016</td>
-<td><blockquote class="first last">
-<div>Neural Network</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Feed forward neural network, including AlexNet, cuDNN layers,
etc.</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
+<td>Neural Network</td>
+<td>Feed forward neural network, including AlexNet, cuDNN layers,Tools</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Recurrent neural network, including GRULayer and BPTT</div></blockquote>
-</td>
-<td>done</td>
+<td>Recurrent neural network, including GRULayer and BPTT</td>
</tr>
<tr class="row-odd"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Model partition and hybrid partition</div></blockquote>
-</td>
-<td>done</td>
+<td>Model partition and hybrid partition</td>
</tr>
<tr class="row-even"><td> </td>
-<td><blockquote class="first last">
-<div>Tools</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Integration with Mesos for resource management</div></blockquote>
-</td>
-<td>done</td>
+<td>Tools</td>
+<td>Integration with Mesos for resource management</td>
</tr>
<tr class="row-odd"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Prepare Docker images for deployment</div></blockquote>
-</td>
-<td>done</td>
+<td>Prepare Docker images for deployment</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Visualization of neural net and debug information</div></blockquote>
-</td>
-<td>done</td>
+<td>Visualization of neural net and debug information</td>
</tr>
<tr class="row-odd"><td> </td>
-<td><blockquote class="first last">
-<div>Binding</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Python binding for major components</div></blockquote>
-</td>
-<td>done</td>
+<td>Binding</td>
+<td>Python binding for major components</td>
</tr>
<tr class="row-even"><td> </td>
-<td><blockquote class="first last">
-<div>GPU</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Single node with multiple GPUs</div></blockquote>
-</td>
-<td>done</td>
+<td>GPU</td>
+<td>Single node with multiple GPUs</td>
</tr>
<tr class="row-odd"><td>0.3 April 2016</td>
-<td><blockquote class="first last">
-<div>GPU</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Multiple nodes, each with multiple GPUs</div></blockquote>
-</td>
-<td>done</td>
+<td>GPU</td>
+<td>Multiple nodes, each with multiple GPUs</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Heterogeneous training using both GPU and CPU [CcT](<a class="reference
external"
href="http://arxiv.org/abs/1504.04343";>http://arxiv.org/abs/1504.04343</a>)</div></blockquote>
-</td>
-<td>done</td>
+<td>Heterogeneous training using both GPU and CPU <a class="reference
external" href="http://arxiv.org/abs/1504.04343";>CcT</a></td>
</tr>
<tr class="row-odd"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Support cuDNN v4</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>done</div></blockquote>
-</td>
-</tr>
-<tr class="row-even"><td> </td>
-<td><blockquote class="first last">
-<div>Installation</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Remove dependency on ZeroMQ, CZMQ, Zookeeper for single node
training</div></blockquote>
-</td>
-<td>done</td>
-</tr>
-<tr class="row-odd"><td> </td>
-<td><blockquote class="first last">
-<div>Updater</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Add new SGD updaters including Adam, AdamMax and
AdaDelta</div></blockquote>
-</td>
-<td>done</td>
-</tr>
-<tr class="row-even"><td> </td>
-<td><blockquote class="first last">
-<div>Binding</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Enhance Python binding for training</div></blockquote>
-</td>
-<td>done</td>
+<td>Support cuDNN v4</td>
+</tr>
+<tr class="row-even"><td> </td>
+<td>Installation</td>
+<td>Remove dependency on ZeroMQ, CZMQ, Zookeeper for single node training</td>
+</tr>
+<tr class="row-odd"><td> </td>
+<td>Updater</td>
+<td>Add new SGD updaters including Adam, AdamMax and AdaDelta</td>
+</tr>
+<tr class="row-even"><td> </td>
+<td>Binding</td>
+<td>Enhance Python binding for training</td>
</tr>
<tr class="row-odd"><td>1.0 Sep 2016</td>
-<td><blockquote class="first last">
-<div>Programming abstraction</div></blockquote>
-</td>
+<td>Programming abstraction</td>
<td>Tensor with linear algebra, neural net and random operations</td>
-<td> </td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
<td>Updater for distributed parameter updating</td>
-<td> </td>
</tr>
<tr class="row-odd"><td> </td>
-<td><blockquote class="first last">
-<div>Hardware</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Use Cuda and Cudnn for Nvidia GPU</div></blockquote>
-</td>
-<td> </td>
+<td>Hardware</td>
+<td>Use Cuda and Cudnn for Nvidia GPU</td>
</tr>
<tr class="row-even"><td> </td>
<td> </td>
-<td><blockquote class="first last">
-<div>Use OpenCL for AMD GPU or other devices</div></blockquote>
-</td>
-<td> </td>
+<td>Use OpenCL for AMD GPU or other devices</td>
</tr>
<tr class="row-odd"><td> </td>
-<td><blockquote class="first last">
-<div>Cross-platform</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>To extend from Linux to MacOS</div></blockquote>
-</td>
-<td> </td>
+<td>Cross-platform</td>
+<td>To extend from Linux to MacOS</td>
</tr>
<tr class="row-even"><td> </td>
-<td><blockquote class="first last">
-<div>Examples</div></blockquote>
-</td>
-<td><blockquote class="first last">
-<div>Speech recognition example</div></blockquote>
-</td>
<td> </td>
+<td>Large image models, e.g., <a class="reference external"
href="https://arxiv.org/pdf/1409.1556.pdf";>VGG</a> and <a class="reference
external" href="http://arxiv.org/abs/1512.03385";>Residual Net</a></td>
+</tr>
+<tr class="row-odd"><td>1.1 Dec 2016</td>
+<td>Model Zoo</td>
+<td>Health-care models and popular image models</td>
+</tr>
+<tr class="row-even"><td> </td>
+<td>Caffe converter</td>
+<td>Use SINGA to train models configured in caffe proto files</td>
</tr>
<tr class="row-odd"><td> </td>
-<td> </td>
-<td>Large image models, e.g., [VGG](<a class="reference external"
href="https://arxiv.org/pdf/1409.1556.pdf";>https://arxiv.org/pdf/1409.1556.pdf</a>)
and [Residual Net](<a class="reference external"
href="http://arxiv.org/abs/1512.03385";>http://arxiv.org/abs/1512.03385</a>)</td>
-<td> </td>
+<td>Memory optimization</td>
+<td>Replace CNMEM with new memory pool to reduce memory footprint</td>
+</tr>
+<tr class="row-even"><td> </td>
+<td>Distributed training</td>
+<td>Migrate distributed training frameworks from V0.3</td>
+</tr>
+<tr class="row-odd"><td> </td>
+<td>Compilation and installation</td>
+<td>Windows suppport</td>
</tr>
-<tr class="row-even"><td>1.1 Dec 2016</td>
+<tr class="row-even"><td> </td>
<td> </td>
+<td>Simplify the installation by compiling protobuf and openblas together with
SINGA</td>
+</tr>
+<tr class="row-odd"><td> </td>
<td> </td>
+<td>Build python wheel automatically using Jenkins</td>
+</tr>
+<tr class="row-even"><td> </td>
<td> </td>
+<td>Deploy SINGA programs on Android phones for prediction tasks</td>
</tr>
</tbody>
</table>
Modified: incubator/singa/site/trunk/en/docs/installation.html
URL:
http://svn.apache.org/viewvc/incubator/singa/site/trunk/en/docs/installation.html?rev=1763496&r1=1763495&r2=1763496&view=diff
==============================================================================
--- incubator/singa/site/trunk/en/docs/installation.html (original)
+++ incubator/singa/site/trunk/en/docs/installation.html Thu Oct 6 03:20:51
2016
@@ -260,6 +260,7 @@ $ export CMAKE_LIBRARY_PATH=/usr/local/o
<div class="section" id="pip-and-anaconda-for-pysinga">
<span id="pip-and-anaconda-for-pysinga"></span><h3>pip and anaconda for
PySINGA<a class="headerlink" href="#pip-and-anaconda-for-pysinga"
title="Permalink to this headline">¶</a></h3>
<p>pip and anaconda could be used to install python packages, e.g. numpy.
+Python virtual environment is recommended to run PySINGA.
To use pip with virtual environment,</p>
<div class="highlight-default"><div class="highlight"><pre># install virtualenv
$ pip install virtualenv
@@ -407,6 +408,12 @@ Remember to add its directory to <code c
<div class="section" id="faq">
<span id="faq"></span><h2>FAQ<a class="headerlink" href="#faq"
title="Permalink to this headline">¶</a></h2>
<ul>
+<li><p class="first">Q: Error from ‘import singa’ using PySINGA
installed from wheel.</p>
+<p>A: Please check the detailed error from <code class="docutils
literal"><span class="pre">python</span> <span class="pre">-c</span> <span
class="pre">"from</span> <span class="pre">singa</span> <span
class="pre">import</span> <span class="pre">_singa_wrap"</span></code>.
Sometimes it is
+caused by the dependent libraries, e.g. there are multiple versions of
protobuf.</p>
+</li>
+</ul>
+<ul>
<li><p class="first">Q: Error from running <code class="docutils
literal"><span class="pre">cmake</span> <span class="pre">..</span></code>,
which cannot find the dependent libraries.</p>
<p>A: If you haven’t installed the libraries, please install them. If
you installed
the libraries in a folder that is outside of the system folder, e.g.
/usr/local,
@@ -463,7 +470,7 @@ e.g.,</p>
</div>
</li>
<li><p class="first">Q: When I build protocol buffer, it reports that
GLIBC++_3.4.20 not found in /usr/lib64/libstdc++.so.6.</p>
-<p>A9: This means the linker found libstdc++.so.6 but that library
+<p>A: This means the linker found libstdc++.so.6 but that library
belongs to an older version of GCC than was used to compile and link the
program. The program depends on code defined in
the newer libstdc++ that belongs to the newer version of GCC, so the linker
