Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_0.1.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_0.1.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_0.1.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_0.1.0.html Sat Jun 29 14:42:24 2019 @@ -205,94 +205,102 @@ designed with an intuitive programming m deep learning models.</p> <p>This release includes following features:</p> <ul class="simple"> -<li>Job management<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-3";>SINGA-3</a> Use Zookeeper to check stopping (finish) time of the system</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-16";>SINGA-16</a> Runtime Process id Management</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-25";>SINGA-25</a> Setup glog output path</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-26";>SINGA-26</a> Run distributed training in a single command</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-30";>SINGA-30</a> Enhance easy-to-use feature and support concurrent jobs</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-33";>SINGA-33</a> Automatically launch a number of processes in the cluster</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-34";>SINGA-34</a> Support external zookeeper service</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-38";>SINGA-38</a> Support concurrent jobs</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-39";>SINGA-39</a> Avoid ssh in scripts for single node environment</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-43";>SINGA-43</a> Remove Job-related output from workspace</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-56";>SINGA-56</a> No automatic launching of zookeeper service</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-73";>SINGA-73</a> Refine the selection of available hosts from host list</li> -</ul> -</li> -<li>Installation with GNU Auto tool<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-4";>SINGA-4</a> Refine thirdparty-dependency installation</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-13";>SINGA-13</a> Separate intermediate files of compilation from source files</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-17";>SINGA-17</a> Add root permission within thirdparty/install.</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-27";>SINGA-27</a> Generate python modules for proto objects</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-53";>SINGA-53</a> Add lmdb compiling options</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-62";>SINGA-62</a> Remove building scrips and auxiliary files</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-67";>SINGA-67</a> Add singatest into build targets</li> -</ul> -</li> -<li>Distributed training<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-7";>SINGA-7</a> Implement shared memory Hogwild algorithm</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-8";>SINGA-8</a> Implement distributed Hogwild</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-19";>SINGA-19</a> Slice large Param objects for load-balance</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-29";>SINGA-29</a> Update NeuralNet class to enable layer partition type customization</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-24";>SINGA-24</a> Implement Downpour training framework</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-32";>SINGA-32</a> Implement AllReduce training framework</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-57";>SINGA-57</a> Improve Distributed Hogwild</li> -</ul> -</li> -<li>Training algorithms for different model categories<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-9";>SINGA-9</a> Add Support for Restricted Boltzman Machine (RBM) model</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-10";>SINGA-10</a> Add Support for Recurrent Neural Networks (RNN)</li> -</ul> -</li> -<li>Checkpoint and restore<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-12";>SINGA-12</a> Support Checkpoint and Restore</li> -</ul> -</li> -<li>Unit test<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-64";>SINGA-64</a> Add the test module for utils/common</li> -</ul> -</li> -<li>Programming model<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-36";>SINGA-36</a> Refactor job configuration, driver program and scripts</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-37";>SINGA-37</a> Enable users to set parameter sharing in model configuration</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-54";>SINGA-54</a> Refactor job configuration to move fields in ModelProto out</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-55";>SINGA-55</a> Refactor main.cc and singa.h</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-61";>SINGA-61</a> Support user defined classes</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-65";>SINGA-65</a> Add an example of writing user-defined layers</li> -</ul> -</li> -<li>Other features<ul> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-6";>SINGA-6</a> Implement thread-safe singleton</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-18";>SINGA-18</a> Update API for displaying performance metric</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-77";>SINGA-77</a> Integrate with Apache RAT</li> +<li><p>Job management</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-3";>SINGA-3</a> Use Zookeeper to check stopping (finish) time of the system</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-16";>SINGA-16</a> Runtime Process id Management</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-25";>SINGA-25</a> Setup glog output path</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-26";>SINGA-26</a> Run distributed training in a single command</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-30";>SINGA-30</a> Enhance easy-to-use feature and support concurrent jobs</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-33";>SINGA-33</a> Automatically launch a number of processes in the cluster</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-34";>SINGA-34</a> Support external zookeeper service</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-38";>SINGA-38</a> Support concurrent jobs</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-39";>SINGA-39</a> Avoid ssh in scripts for single node environment</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-43";>SINGA-43</a> Remove Job-related output from workspace</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-56";>SINGA-56</a> No automatic launching of zookeeper service</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-73";>SINGA-73</a> Refine the selection of available hosts from host list</p></li> +</ul> +</li> +<li><p>Installation with GNU Auto tool</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-4";>SINGA-4</a> Refine thirdparty-dependency installation</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-13";>SINGA-13</a> Separate intermediate files of compilation from source files</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-17";>SINGA-17</a> Add root permission within thirdparty/install.</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-27";>SINGA-27</a> Generate python modules for proto objects</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-53";>SINGA-53</a> Add lmdb compiling options</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-62";>SINGA-62</a> Remove building scrips and auxiliary files</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-67";>SINGA-67</a> Add singatest into build targets</p></li> +</ul> +</li> +<li><p>Distributed training</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-7";>SINGA-7</a> Implement shared memory Hogwild algorithm</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-8";>SINGA-8</a> Implement distributed Hogwild</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-19";>SINGA-19</a> Slice large Param objects for load-balance</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-29";>SINGA-29</a> Update NeuralNet class to enable layer partition type customization</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-24";>SINGA-24</a> Implement Downpour training framework</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-32";>SINGA-32</a> Implement AllReduce training framework</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-57";>SINGA-57</a> Improve Distributed Hogwild</p></li> +</ul> +</li> +<li><p>Training algorithms for different model categories</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-9";>SINGA-9</a> Add Support for Restricted Boltzman Machine (RBM) model</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-10";>SINGA-10</a> Add Support for Recurrent Neural Networks (RNN)</p></li> +</ul> +</li> +<li><p>Checkpoint and restore</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-12";>SINGA-12</a> Support Checkpoint and Restore</p></li> +</ul> +</li> +<li><p>Unit test</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-64";>SINGA-64</a> Add the test module for utils/common</p></li> +</ul> +</li> +<li><p>Programming model</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-36";>SINGA-36</a> Refactor job configuration, driver program and scripts</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-37";>SINGA-37</a> Enable users to set parameter sharing in model configuration</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-54";>SINGA-54</a> Refactor job configuration to move fields in ModelProto out</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-55";>SINGA-55</a> Refactor main.cc and singa.h</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-61";>SINGA-61</a> Support user defined classes</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-65";>SINGA-65</a> Add an example of writing user-defined layers</p></li> +</ul> +</li> +<li><p>Other features</p> +<ul> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-6";>SINGA-6</a> Implement thread-safe singleton</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-18";>SINGA-18</a> Update API for displaying performance metric</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-77";>SINGA-77</a> Integrate with Apache RAT</p></li> </ul> </li> </ul> <p>Some bugs are fixed during the development of this release</p> <ul class="simple"> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-2";>SINGA-2</a> Check failed: zsock_connect</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-5";>SINGA-5</a> Server early terminate when zookeeper singa folder is not initially empty</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-15";>SINGA-15</a> Fixg a bug from ConnectStub function which gets stuck for connecting layer_dealer_</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-22";>SINGA-22</a> Cannot find openblas library when it is installed in default path</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-23";>SINGA-23</a> Libtool version mismatch error.</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-28";>SINGA-28</a> Fix a bug from topology sort of Graph</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-42";>SINGA-42</a> Issue when loading checkpoints</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-44";>SINGA-44</a> A bug when reseting metric values</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-46";>SINGA-46</a> Fix a bug in updater.cc to scale the gradients</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-47";>SINGA-47</a> Fix a bug in data layers that leads to out-of-memory when group size is too large</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-48";>SINGA-48</a> Fix a bug in trainer.cc that assigns the same NeuralNet instance to workers from diff groups</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-49";>SINGA-49</a> Fix a bug in HandlePutMsg func that sets param fields to invalid values</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-66";>SINGA-66</a> Fix bugs in Worker::RunOneBatch function and ClusterProto</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-79";>SINGA-79</a> Fix bug in singatool that can not parse -conf flag</li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-2";>SINGA-2</a> Check failed: zsock_connect</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-5";>SINGA-5</a> Server early terminate when zookeeper singa folder is not initially empty</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-15";>SINGA-15</a> Fixg a bug from ConnectStub function which gets stuck for connecting layer_dealer_</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-22";>SINGA-22</a> Cannot find openblas library when it is installed in default path</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-23";>SINGA-23</a> Libtool version mismatch error.</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-28";>SINGA-28</a> Fix a bug from topology sort of Graph</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-42";>SINGA-42</a> Issue when loading checkpoints</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-44";>SINGA-44</a> A bug when reseting metric values</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-46";>SINGA-46</a> Fix a bug in updater.cc to scale the gradients</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-47";>SINGA-47</a> Fix a bug in data layers that leads to out-of-memory when group size is too large</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-48";>SINGA-48</a> Fix a bug in trainer.cc that assigns the same NeuralNet instance to workers from diff groups</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-49";>SINGA-49</a> Fix a bug in HandlePutMsg func that sets param fields to invalid values</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-66";>SINGA-66</a> Fix bugs in Worker::RunOneBatch function and ClusterProto</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-79";>SINGA-79</a> Fix bug in singatool that can not parse -conf flag</p></li> </ul> <p>Features planned for the next release</p> <ul class="simple"> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-11";>SINGA-11</a> Start SINGA using Mesos</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-31";>SINGA-31</a> Extend Blob to support xpu (cpu or gpu)</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-35";>SINGA-35</a> Add random number generators</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-40";>SINGA-40</a> Support sparse Param update</li> -<li><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-41";>SINGA-41</a> Support single node single GPU training</li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-11";>SINGA-11</a> Start SINGA using Mesos</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-31";>SINGA-31</a> Extend Blob to support xpu (cpu or gpu)</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-35";>SINGA-35</a> Add random number generators</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-40";>SINGA-40</a> Support sparse Param update</p></li> +<li><p><a class="reference external" href="https://issues.apache.org/jira/browse/SINGA-41";>SINGA-41</a> Support single node single GPU training</p></li> </ul> </div>
Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_0.2.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_0.2.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_0.2.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_0.2.0.html Sat Jun 29 14:42:24 2019 @@ -206,76 +206,86 @@ programming model based on the layer abs of popular deep learning models.</p> <p>This release includes the following <strong>major features</strong>:</p> <ul class="simple"> -<li><a class="reference external" href="../docs/gpu.html">Training on GPU</a> enables training of complex models on a single node with multiple GPU cards.</li> -<li><a class="reference external" href="../docs/hybrid.html">Hybrid neural net partitioning</a> supports data and model parallelism at the same time.</li> -<li><a class="reference external" href="../docs/python.html">Python wrapper</a> makes it easy to configure the job, including neural net and SGD algorithm.</li> -<li><a class="reference external" href="../docs/general-rnn.html">RNN model and BPTT algorithm</a> are implemented to support applications based on RNN models, e.g., GRU.</li> -<li><a class="reference external" href="../docs/distributed-training">Cloud software integration</a> includes Mesos, Docker and HDFS.</li> +<li><p><a class="reference external" href="../docs/gpu.html">Training on GPU</a> enables training of complex models on a single node with multiple GPU cards.</p></li> +<li><p><a class="reference external" href="../docs/hybrid.html">Hybrid neural net partitioning</a> supports data and model parallelism at the same time.</p></li> +<li><p><a class="reference external" href="../docs/python.html">Python wrapper</a> makes it easy to configure the job, including neural net and SGD algorithm.</p></li> +<li><p><a class="reference external" href="../docs/general-rnn.html">RNN model and BPTT algorithm</a> are implemented to support applications based on RNN models, e.g., GRU.</p></li> +<li><p><a class="reference external" href="../docs/distributed-training">Cloud software integration</a> includes Mesos, Docker and HDFS.</p></li> </ul> <p><strong>More details</strong> are listed as follows,</p> <ul class="simple"> -<li>Programming model<ul> -<li>[SINGA-80] New Blob Level and Address Level Math Operation Interface</li> -<li>[SINGA-82] Refactor input layers using data store abstraction</li> -<li>[SINGA-87] Replace exclude field to include field for layer configuration</li> -<li>[SINGA-110] Add Layer member datavec_ and gradvec_</li> -<li>[SINGA-120] Implemented GRU and BPTT (BPTTWorker)</li> -</ul> -</li> -<li>Neuralnet layers<ul> -<li>[SINGA-91] Add SoftmaxLayer and ArgSortLayer</li> -<li>[SINGA-106] Add dummy layer for test purpose</li> -<li>[SINGA-120] Implemented GRU and BPTT (GRULayer and OneHotLayer)</li> -</ul> -</li> -<li>GPU training support<ul> -<li>[SINGA-100] Implement layers using CUDNN for GPU training</li> -<li>[SINGA-104] Add Context Class</li> -<li>[SINGA-105] Update GUN make files for compiling cuda related code</li> -<li>[SINGA-98] Add Support for AlexNet ImageNet Classification Model</li> -</ul> -</li> -<li>Model/Hybrid partition<ul> -<li>[SINGA-109] Refine bridge layers</li> -<li>[SINGA-111] Add slice, concate and split layers</li> -<li>[SINGA-113] Model/Hybrid Partition Support</li> -</ul> -</li> -<li>Python binding<ul> -<li>[SINGA-108] Add Python wrapper to singa</li> -</ul> -</li> -<li>Predict-only mode<ul> -<li>[SINGA-85] Add functions for extracting features and test new data</li> -</ul> -</li> -<li>Integrate with third-party tools<ul> -<li>[SINGA-11] Start SINGA on Apache Mesos</li> -<li>[SINGA-78] Use Doxygen to generate documentation</li> -<li>[SINGA-89] Add Docker support</li> -</ul> -</li> -<li>Unit test<ul> -<li>[SINGA-95] Add make test after building</li> -</ul> -</li> -<li>Other improvment<ul> -<li>[SINGA-84] Header Files Rearrange</li> -<li>[SINGA-93] Remove the asterisk in the log tcp://169.254.12.152:*:49152</li> -<li>[SINGA-94] Move call to google::InitGoogleLogging() from Driver::Init() to main()</li> -<li>[SINGA-96] Add Momentum to Cifar10 Example</li> -<li>[SINGA-101] Add ll (ls -l) command in .bashrc file when using docker</li> -<li>[SINGA-114] Remove short logs in tmp directory</li> -<li>[SINGA-115] Print layer debug information in the neural net graph file</li> -<li>[SINGA-118] Make protobuf LayerType field id easy to assign</li> -<li>[SIGNA-97] Add HDFS Store</li> -</ul> -</li> -<li>Bugs fixed<ul> -<li>[SINGA-85] Fix compilation errors in examples</li> -<li>[SINGA-90] Miscellaneous trivial bug fixes</li> -<li>[SINGA-107] Error from loading pre-trained params for training stacked RBMs</li> -<li>[SINGA-116] Fix a bug in InnerProductLayer caused by weight matrix sharing</li> +<li><p>Programming model</p> +<ul> +<li><p>[SINGA-80] New Blob Level and Address Level Math Operation Interface</p></li> +<li><p>[SINGA-82] Refactor input layers using data store abstraction</p></li> +<li><p>[SINGA-87] Replace exclude field to include field for layer configuration</p></li> +<li><p>[SINGA-110] Add Layer member datavec_ and gradvec_</p></li> +<li><p>[SINGA-120] Implemented GRU and BPTT (BPTTWorker)</p></li> +</ul> +</li> +<li><p>Neuralnet layers</p> +<ul> +<li><p>[SINGA-91] Add SoftmaxLayer and ArgSortLayer</p></li> +<li><p>[SINGA-106] Add dummy layer for test purpose</p></li> +<li><p>[SINGA-120] Implemented GRU and BPTT (GRULayer and OneHotLayer)</p></li> +</ul> +</li> +<li><p>GPU training support</p> +<ul> +<li><p>[SINGA-100] Implement layers using CUDNN for GPU training</p></li> +<li><p>[SINGA-104] Add Context Class</p></li> +<li><p>[SINGA-105] Update GUN make files for compiling cuda related code</p></li> +<li><p>[SINGA-98] Add Support for AlexNet ImageNet Classification Model</p></li> +</ul> +</li> +<li><p>Model/Hybrid partition</p> +<ul> +<li><p>[SINGA-109] Refine bridge layers</p></li> +<li><p>[SINGA-111] Add slice, concate and split layers</p></li> +<li><p>[SINGA-113] Model/Hybrid Partition Support</p></li> +</ul> +</li> +<li><p>Python binding</p> +<ul> +<li><p>[SINGA-108] Add Python wrapper to singa</p></li> +</ul> +</li> +<li><p>Predict-only mode</p> +<ul> +<li><p>[SINGA-85] Add functions for extracting features and test new data</p></li> +</ul> +</li> +<li><p>Integrate with third-party tools</p> +<ul> +<li><p>[SINGA-11] Start SINGA on Apache Mesos</p></li> +<li><p>[SINGA-78] Use Doxygen to generate documentation</p></li> +<li><p>[SINGA-89] Add Docker support</p></li> +</ul> +</li> +<li><p>Unit test</p> +<ul> +<li><p>[SINGA-95] Add make test after building</p></li> +</ul> +</li> +<li><p>Other improvment</p> +<ul> +<li><p>[SINGA-84] Header Files Rearrange</p></li> +<li><p>[SINGA-93] Remove the asterisk in the log tcp://169.254.12.152:*:49152</p></li> +<li><p>[SINGA-94] Move call to google::InitGoogleLogging() from Driver::Init() to main()</p></li> +<li><p>[SINGA-96] Add Momentum to Cifar10 Example</p></li> +<li><p>[SINGA-101] Add ll (ls -l) command in .bashrc file when using docker</p></li> +<li><p>[SINGA-114] Remove short logs in tmp directory</p></li> +<li><p>[SINGA-115] Print layer debug information in the neural net graph file</p></li> +<li><p>[SINGA-118] Make protobuf LayerType field id easy to assign</p></li> +<li><p>[SIGNA-97] Add HDFS Store</p></li> +</ul> +</li> +<li><p>Bugs fixed</p> +<ul> +<li><p>[SINGA-85] Fix compilation errors in examples</p></li> +<li><p>[SINGA-90] Miscellaneous trivial bug fixes</p></li> +<li><p>[SINGA-107] Error from loading pre-trained params for training stacked RBMs</p></li> +<li><p>[SINGA-116] Fix a bug in InnerProductLayer caused by weight matrix sharing</p></li> </ul> </li> </ul> Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_0.3.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_0.3.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_0.3.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_0.3.0.html Sat Jun 29 14:42:24 2019 @@ -206,34 +206,39 @@ programming model based on the layer abs of popular deep learning models.</p> <p>This release includes following features:</p> <ul class="simple"> -<li>GPU Support<ul> -<li>[SINGA-131] Implement and optimize hybrid training using both CPU and GPU</li> -<li>[SINGA-136] Support cuDNN v4</li> -<li>[SINGA-134] Extend SINGA to run over a GPU cluster</li> -<li>[SINGA-157] Change the priority of cudnn library and install libsingagpu.so</li> +<li><p>GPU Support</p> +<ul> +<li><p>[SINGA-131] Implement and optimize hybrid training using both CPU and GPU</p></li> +<li><p>[SINGA-136] Support cuDNN v4</p></li> +<li><p>[SINGA-134] Extend SINGA to run over a GPU cluster</p></li> +<li><p>[SINGA-157] Change the priority of cudnn library and install libsingagpu.so</p></li> </ul> </li> -<li>Remove Dependences<ul> -<li>[SINGA-156] Remove the dependency on ZMQ for single process training</li> -<li>[SINGA-155] Remove zookeeper for single-process training</li> +<li><p>Remove Dependences</p> +<ul> +<li><p>[SINGA-156] Remove the dependency on ZMQ for single process training</p></li> +<li><p>[SINGA-155] Remove zookeeper for single-process training</p></li> </ul> </li> -<li>Python Binding<ul> -<li>[SINGA-126] Python Binding for Interactive Training</li> +<li><p>Python Binding</p> +<ul> +<li><p>[SINGA-126] Python Binding for Interactive Training</p></li> </ul> </li> -<li>Other Improvements<ul> -<li>[SINGA-80] New Blob Level and Address Level Math Operation Interface</li> -<li>[SINGA-130] Data Prefetching</li> -<li>[SINGA-145] New SGD based optimization Updaters: AdaDelta, Adam, AdamMax</li> +<li><p>Other Improvements</p> +<ul> +<li><p>[SINGA-80] New Blob Level and Address Level Math Operation Interface</p></li> +<li><p>[SINGA-130] Data Prefetching</p></li> +<li><p>[SINGA-145] New SGD based optimization Updaters: AdaDelta, Adam, AdamMax</p></li> </ul> </li> -<li>Bugs Fixed<ul> -<li>[SINGA-148] Race condition between Worker threads and Driver</li> -<li>[SINGA-150] Mesos Docker container failed</li> -<li>[SIGNA-141] Undesired Hash collision when locating process id to workerâ¦</li> -<li>[SINGA-149] Docker build fail</li> -<li>[SINGA-143] The compilation cannot detect libsingagpu.so file</li> +<li><p>Bugs Fixed</p> +<ul> +<li><p>[SINGA-148] Race condition between Worker threads and Driver</p></li> +<li><p>[SINGA-150] Mesos Docker container failed</p></li> +<li><p>[SIGNA-141] Undesired Hash collision when locating process id to workerâ¦</p></li> +<li><p>[SINGA-149] Docker build fail</p></li> +<li><p>[SINGA-143] The compilation cannot detect libsingagpu.so file</p></li> </ul> </li> </ul> Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_1.0.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_1.0.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_1.0.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_1.0.0.html Sat Jun 29 14:42:24 2019 @@ -206,87 +206,93 @@ programming model based on the layer abs of popular deep learning models.</p> <p>This release includes following features:</p> <ul class="simple"> -<li>Core abstractions including Tensor and Device<ul> -<li>[SINGA-207] Update Tensor functions for matrices</li> -<li>[SINGA-205] Enable slice and concatenate operations for Tensor objects</li> -<li>[SINGA-197] Add CNMem as a submodule in lib/</li> -<li>[SINGA-196] Rename class Blob to Block</li> -<li>[SINGA-194] Add a Platform singleton</li> -<li>[SINGA-175] Add memory management APIs and implement a subclass using CNMeM</li> -<li>[SINGA-173] OpenCL Implementation</li> -<li>[SINGA-171] Create CppDevice and CudaDevice</li> -<li>[SINGA-168] Implement Cpp Math functions APIs</li> -<li>[SINGA-162] Overview of features for V1.x</li> -<li>[SINGA-165] Add cross-platform timer API to singa</li> -<li>[SINGA-167] Add Tensor Math function APIs</li> -<li>[SINGA-166] light built-in logging for making glog optional</li> -<li>[SINGA-164] Add the base Tensor class</li> -</ul> -</li> -<li>IO components for file read/write, network and data pre-processing<ul> -<li>[SINGA-233] New communication interface</li> -<li>[SINGA-215] Implement Image Transformation for Image Pre-processing</li> -<li>[SINGA-214] Add LMDBReader and LMDBWriter for LMDB</li> -<li>[SINGA-213] Implement Encoder and Decoder for CSV</li> -<li>[SINGA-211] Add TextFileReader and TextFileWriter for CSV files</li> -<li>[SINGA-210] Enable checkpoint and resume for v1.0</li> -<li>[SINGA-208] Add DataIter base class and a simple implementation</li> -<li>[SINGA-203] Add OpenCV detection for cmake compilation</li> -<li>[SINGA-202] Add reader and writer for binary file</li> -<li>[SINGA-200] Implement Encoder and Decoder for data pre-processing</li> -</ul> -</li> -<li>Module components including layer classes, training algorithms and Python binding<ul> -<li>[SINGA-235] Unify the engines for cudnn and singa layers</li> -<li>[SINGA-230] OpenCL Convolution layer and Pooling layer</li> -<li>[SINGA-222] Fixed bugs in IO</li> -<li>[SINGA-218] Implementation for RNN CUDNN version</li> -<li>[SINGA-204] Support the training of feed-forward neural nets</li> -<li>[SINGA-199] Implement Python classes for SGD optimizers</li> -<li>[SINGA-198] Change Layer::Setup API to include input Tensor shapes</li> -<li>[SINGA-193] Add Python layers</li> -<li>[SINGA-192] Implement optimization algorithms for SINGA v1 (nesterove, adagrad, rmsprop)</li> -<li>[SINGA-191] Add âautotuneâ for CudnnConvolution Layer</li> -<li>[SINGA-190] Add prelu layer and flatten layer</li> -<li>[SINGA-189] Generate python outputs of proto files</li> -<li>[SINGA-188] Add Dense layer</li> -<li>[SINGA-187] Add popular parameter initialization methods</li> -<li>[SINGA-186] Create Python Tensor class</li> -<li>[SINGA-184] Add Cross Entropy loss computation</li> -<li>[SINGA-183] Add the base classes for optimizer, constraint and regularizer</li> -<li>[SINGA-180] Add Activation layer and Softmax layer</li> -<li>[SINGA-178] Add Convolution layer and Pooling layer</li> -<li>[SINGA-176] Add loss and metric base classes</li> -<li>[SINGA-174] Add Batch Normalization layer and Local Response Nomalization layer.</li> -<li>[SINGA-170] Add Dropout layer and CudnnDropout layer.</li> -<li>[SINGA-169] Add base Layer class for V1.0</li> -</ul> -</li> -<li>Examples<ul> -<li>[SINGA-232] Alexnet on Imagenet</li> -<li>[SINGA-231] Batchnormlized VGG model for cifar-10</li> -<li>[SINGA-228] Add Cpp Version of Convolution and Pooling layer</li> -<li>[SINGA-227] Add Split and Merge Layer and add ResNet Implementation</li> -</ul> -</li> -<li>Documentation<ul> -<li>[SINGA-239] Transfer documentation files of v0.3.0 to github</li> -<li>[SINGA-238] RBM on mnist</li> -<li>[SINGA-225] Documentation for installation and Cifar10 example</li> -<li>[SINGA-223] Use Sphinx to create the website</li> -</ul> -</li> -<li>Tools for compilation and some utility code<ul> -<li>[SINGA-229] Complete install targets</li> -<li>[SINGA-221] Support for Travis-CI</li> -<li>[SINGA-217] build python package with setup.py</li> -<li>[SINGA-216] add jenkins for CI support</li> -<li>[SINGA-212] Disable the compilation of libcnmem if USE_CUDA is OFF</li> -<li>[SINGA-195] Channel for sending training statistics</li> -<li>[SINGA-185] Add CBLAS and GLOG detection for singav1</li> -<li>[SINGA-181] Add NVCC supporting for .cu files</li> -<li>[SINGA-177] Add fully cmake supporting for the compilation of singa_v1</li> -<li>[SINGA-172] Add CMake supporting for Cuda and Cudnn libs</li> +<li><p>Core abstractions including Tensor and Device</p> +<ul> +<li><p>[SINGA-207] Update Tensor functions for matrices</p></li> +<li><p>[SINGA-205] Enable slice and concatenate operations for Tensor objects</p></li> +<li><p>[SINGA-197] Add CNMem as a submodule in lib/</p></li> +<li><p>[SINGA-196] Rename class Blob to Block</p></li> +<li><p>[SINGA-194] Add a Platform singleton</p></li> +<li><p>[SINGA-175] Add memory management APIs and implement a subclass using CNMeM</p></li> +<li><p>[SINGA-173] OpenCL Implementation</p></li> +<li><p>[SINGA-171] Create CppDevice and CudaDevice</p></li> +<li><p>[SINGA-168] Implement Cpp Math functions APIs</p></li> +<li><p>[SINGA-162] Overview of features for V1.x</p></li> +<li><p>[SINGA-165] Add cross-platform timer API to singa</p></li> +<li><p>[SINGA-167] Add Tensor Math function APIs</p></li> +<li><p>[SINGA-166] light built-in logging for making glog optional</p></li> +<li><p>[SINGA-164] Add the base Tensor class</p></li> +</ul> +</li> +<li><p>IO components for file read/write, network and data pre-processing</p> +<ul> +<li><p>[SINGA-233] New communication interface</p></li> +<li><p>[SINGA-215] Implement Image Transformation for Image Pre-processing</p></li> +<li><p>[SINGA-214] Add LMDBReader and LMDBWriter for LMDB</p></li> +<li><p>[SINGA-213] Implement Encoder and Decoder for CSV</p></li> +<li><p>[SINGA-211] Add TextFileReader and TextFileWriter for CSV files</p></li> +<li><p>[SINGA-210] Enable checkpoint and resume for v1.0</p></li> +<li><p>[SINGA-208] Add DataIter base class and a simple implementation</p></li> +<li><p>[SINGA-203] Add OpenCV detection for cmake compilation</p></li> +<li><p>[SINGA-202] Add reader and writer for binary file</p></li> +<li><p>[SINGA-200] Implement Encoder and Decoder for data pre-processing</p></li> +</ul> +</li> +<li><p>Module components including layer classes, training algorithms and Python binding</p> +<ul> +<li><p>[SINGA-235] Unify the engines for cudnn and singa layers</p></li> +<li><p>[SINGA-230] OpenCL Convolution layer and Pooling layer</p></li> +<li><p>[SINGA-222] Fixed bugs in IO</p></li> +<li><p>[SINGA-218] Implementation for RNN CUDNN version</p></li> +<li><p>[SINGA-204] Support the training of feed-forward neural nets</p></li> +<li><p>[SINGA-199] Implement Python classes for SGD optimizers</p></li> +<li><p>[SINGA-198] Change Layer::Setup API to include input Tensor shapes</p></li> +<li><p>[SINGA-193] Add Python layers</p></li> +<li><p>[SINGA-192] Implement optimization algorithms for SINGA v1 (nesterove, adagrad, rmsprop)</p></li> +<li><p>[SINGA-191] Add âautotuneâ for CudnnConvolution Layer</p></li> +<li><p>[SINGA-190] Add prelu layer and flatten layer</p></li> +<li><p>[SINGA-189] Generate python outputs of proto files</p></li> +<li><p>[SINGA-188] Add Dense layer</p></li> +<li><p>[SINGA-187] Add popular parameter initialization methods</p></li> +<li><p>[SINGA-186] Create Python Tensor class</p></li> +<li><p>[SINGA-184] Add Cross Entropy loss computation</p></li> +<li><p>[SINGA-183] Add the base classes for optimizer, constraint and regularizer</p></li> +<li><p>[SINGA-180] Add Activation layer and Softmax layer</p></li> +<li><p>[SINGA-178] Add Convolution layer and Pooling layer</p></li> +<li><p>[SINGA-176] Add loss and metric base classes</p></li> +<li><p>[SINGA-174] Add Batch Normalization layer and Local Response Nomalization layer.</p></li> +<li><p>[SINGA-170] Add Dropout layer and CudnnDropout layer.</p></li> +<li><p>[SINGA-169] Add base Layer class for V1.0</p></li> +</ul> +</li> +<li><p>Examples</p> +<ul> +<li><p>[SINGA-232] Alexnet on Imagenet</p></li> +<li><p>[SINGA-231] Batchnormlized VGG model for cifar-10</p></li> +<li><p>[SINGA-228] Add Cpp Version of Convolution and Pooling layer</p></li> +<li><p>[SINGA-227] Add Split and Merge Layer and add ResNet Implementation</p></li> +</ul> +</li> +<li><p>Documentation</p> +<ul> +<li><p>[SINGA-239] Transfer documentation files of v0.3.0 to github</p></li> +<li><p>[SINGA-238] RBM on mnist</p></li> +<li><p>[SINGA-225] Documentation for installation and Cifar10 example</p></li> +<li><p>[SINGA-223] Use Sphinx to create the website</p></li> +</ul> +</li> +<li><p>Tools for compilation and some utility code</p> +<ul> +<li><p>[SINGA-229] Complete install targets</p></li> +<li><p>[SINGA-221] Support for Travis-CI</p></li> +<li><p>[SINGA-217] build python package with setup.py</p></li> +<li><p>[SINGA-216] add jenkins for CI support</p></li> +<li><p>[SINGA-212] Disable the compilation of libcnmem if USE_CUDA is OFF</p></li> +<li><p>[SINGA-195] Channel for sending training statistics</p></li> +<li><p>[SINGA-185] Add CBLAS and GLOG detection for singav1</p></li> +<li><p>[SINGA-181] Add NVCC supporting for .cu files</p></li> +<li><p>[SINGA-177] Add fully cmake supporting for the compilation of singa_v1</p></li> +<li><p>[SINGA-172] Add CMake supporting for Cuda and Cudnn libs</p></li> </ul> </li> </ul> Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_1.1.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_1.1.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_1.1.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_1.1.0.html Sat Jun 29 14:42:24 2019 @@ -204,48 +204,53 @@ learning models over large datasets.</p> <p>This release includes following features:</p> <ul class="simple"> -<li>Core components<ul> -<li>[SINGA-296] Add sign and to_host function for pysinga tensor module</li> +<li><p>Core components</p> +<ul> +<li><p>[SINGA-296] Add sign and to_host function for pysinga tensor module</p></li> </ul> </li> -<li>Model components<ul> -<li>[SINGA-254] Implement Adam for V1</li> -<li>[SINGA-264] Extend the FeedForwardNet to accept multiple inputs</li> -<li>[SINGA-267] Add spatial mode in batch normalization layer</li> -<li>[SINGA-271] Add Concat and Slice layers</li> -<li>[SINGA-275] Add Cross Entropy Loss for multiple labels</li> -<li>[SINGA-278] Convert trained caffe parameters to singa</li> -<li>[SINGA-287] Add memory size check for cudnn convolution</li> +<li><p>Model components</p> +<ul> +<li><p>[SINGA-254] Implement Adam for V1</p></li> +<li><p>[SINGA-264] Extend the FeedForwardNet to accept multiple inputs</p></li> +<li><p>[SINGA-267] Add spatial mode in batch normalization layer</p></li> +<li><p>[SINGA-271] Add Concat and Slice layers</p></li> +<li><p>[SINGA-275] Add Cross Entropy Loss for multiple labels</p></li> +<li><p>[SINGA-278] Convert trained caffe parameters to singa</p></li> +<li><p>[SINGA-287] Add memory size check for cudnn convolution</p></li> </ul> </li> -<li>Utility functions and CI<ul> -<li>[SINGA-242] Compile all source files into a single library.</li> -<li>[SINGA-244] Separating swig interface and python binding files</li> -<li>[SINGA-246] Imgtool for image augmentation</li> -<li>[SINGA-247] Add windows support for singa</li> -<li>[SINGA-251] Implement image loader for pysinga</li> -<li>[SINGA-252] Use the snapshot methods to dump and load models for pysinga</li> -<li>[SINGA-255] Compile mandatory dependent libaries together with SINGA code</li> -<li>[SINGA-259] Add maven pom file for building java classes</li> -<li>[SINGA-261] Add version ID into the checkpoint files</li> -<li>[SINGA-266] Add Rafiki python toolkits</li> -<li>[SINGA-273] Improve license and contributions</li> -<li>[SINGA-284] Add python unittest into Jenkins and link static libs into whl file</li> -<li>[SINGA-280] Jenkins CI support</li> -<li>[SINGA-288] Publish wheel of PySINGA generated by Jenkins to public servers</li> +<li><p>Utility functions and CI</p> +<ul> +<li><p>[SINGA-242] Compile all source files into a single library.</p></li> +<li><p>[SINGA-244] Separating swig interface and python binding files</p></li> +<li><p>[SINGA-246] Imgtool for image augmentation</p></li> +<li><p>[SINGA-247] Add windows support for singa</p></li> +<li><p>[SINGA-251] Implement image loader for pysinga</p></li> +<li><p>[SINGA-252] Use the snapshot methods to dump and load models for pysinga</p></li> +<li><p>[SINGA-255] Compile mandatory dependent libaries together with SINGA code</p></li> +<li><p>[SINGA-259] Add maven pom file for building java classes</p></li> +<li><p>[SINGA-261] Add version ID into the checkpoint files</p></li> +<li><p>[SINGA-266] Add Rafiki python toolkits</p></li> +<li><p>[SINGA-273] Improve license and contributions</p></li> +<li><p>[SINGA-284] Add python unittest into Jenkins and link static libs into whl file</p></li> +<li><p>[SINGA-280] Jenkins CI support</p></li> +<li><p>[SINGA-288] Publish wheel of PySINGA generated by Jenkins to public servers</p></li> </ul> </li> -<li>Documentation and usability<ul> -<li>[SINGA-263] Create Amazon Machine Image</li> -<li>[SINGA-268] Add IPython notebooks to the documentation</li> -<li>[SINGA-276] Create docker images</li> -<li>[SINGA-289] Update SINGA website automatically using Jenkins</li> -<li>[SINGA-295] Add an example of image classification using GoogleNet</li> +<li><p>Documentation and usability</p> +<ul> +<li><p>[SINGA-263] Create Amazon Machine Image</p></li> +<li><p>[SINGA-268] Add IPython notebooks to the documentation</p></li> +<li><p>[SINGA-276] Create docker images</p></li> +<li><p>[SINGA-289] Update SINGA website automatically using Jenkins</p></li> +<li><p>[SINGA-295] Add an example of image classification using GoogleNet</p></li> </ul> </li> -<li>Bugs fixed<ul> -<li>[SINGA-245] float as the first operand can not multiply with a tensor object</li> -<li>[SINGA-293] Bug from compiling PySINGA on Mac OS X with multiple version of Python</li> +<li><p>Bugs fixed</p> +<ul> +<li><p>[SINGA-245] float as the first operand can not multiply with a tensor object</p></li> +<li><p>[SINGA-293] Bug from compiling PySINGA on Mac OS X with multiple version of Python</p></li> </ul> </li> </ul> Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_1.2.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_1.2.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_1.2.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_1.2.0.html Sat Jun 29 14:42:24 2019 @@ -204,56 +204,61 @@ learning models over large datasets.</p> <p>This release includes following features:</p> <ul class="simple"> -<li>Core components<ul> -<li>[SINGA-290] Upgrade to Python 3</li> -<li>[SINGA-341] Added stride functionality to tensors for CPP</li> -<li>[SINGA-347] Create a function that supports einsum</li> -<li>[SINGA-351] Added stride support and cudnn codes to cuda</li> +<li><p>Core components</p> +<ul> +<li><p>[SINGA-290] Upgrade to Python 3</p></li> +<li><p>[SINGA-341] Added stride functionality to tensors for CPP</p></li> +<li><p>[SINGA-347] Create a function that supports einsum</p></li> +<li><p>[SINGA-351] Added stride support and cudnn codes to cuda</p></li> </ul> </li> -<li>Model components<ul> -<li>[SINGA-300] Add residual networks for imagenet classification</li> -<li>[SINGA-312] Rename layer parameters</li> -<li>[SINGA-313] Add L2 norm layer</li> -<li>[SINGA-315] Reduce memory footprint by Python generator for parameter</li> -<li>[SINGA-316] Add SigmoidCrossEntropy</li> -<li>[SINGA-324] Extend RNN layer to accept variant seq length across batches</li> -<li>[SINGA-326] Add Inception V4 for ImageNet classification</li> -<li>[SINGA-328] Add VGG models for ImageNet classification</li> -<li>[SINGA-329] Support layer freezing during training (fine-tuning)</li> -<li>[SINGA-346] Update cudnn from V5 to V7</li> -<li>[SINGA-349] Create layer operations for autograd</li> -<li>[SINGA-363] Add DenseNet for Imagenet classification</li> +<li><p>Model components</p> +<ul> +<li><p>[SINGA-300] Add residual networks for imagenet classification</p></li> +<li><p>[SINGA-312] Rename layer parameters</p></li> +<li><p>[SINGA-313] Add L2 norm layer</p></li> +<li><p>[SINGA-315] Reduce memory footprint by Python generator for parameter</p></li> +<li><p>[SINGA-316] Add SigmoidCrossEntropy</p></li> +<li><p>[SINGA-324] Extend RNN layer to accept variant seq length across batches</p></li> +<li><p>[SINGA-326] Add Inception V4 for ImageNet classification</p></li> +<li><p>[SINGA-328] Add VGG models for ImageNet classification</p></li> +<li><p>[SINGA-329] Support layer freezing during training (fine-tuning)</p></li> +<li><p>[SINGA-346] Update cudnn from V5 to V7</p></li> +<li><p>[SINGA-349] Create layer operations for autograd</p></li> +<li><p>[SINGA-363] Add DenseNet for Imagenet classification</p></li> </ul> </li> -<li>Utility functions and CI<ul> -<li>[SINGA-274] Improve Debian packaging with CPack</li> -<li>[SINGA-303] Create conda packages</li> -<li>[SINGA-337] Add test cases for code</li> -<li>[SINGA-348] Support autograd MLP Example</li> -<li>[SINGA-345] Update Jenkins and fix bugs in compliation</li> -<li>[SINGA-354] Update travis scripts to use conda-build for all platforms</li> -<li>[SINGA-358] Consolidated RUN steps and cleaned caches in Docker containers</li> -<li>[SINGA-359] Create alias for conda packages</li> +<li><p>Utility functions and CI</p> +<ul> +<li><p>[SINGA-274] Improve Debian packaging with CPack</p></li> +<li><p>[SINGA-303] Create conda packages</p></li> +<li><p>[SINGA-337] Add test cases for code</p></li> +<li><p>[SINGA-348] Support autograd MLP Example</p></li> +<li><p>[SINGA-345] Update Jenkins and fix bugs in compliation</p></li> +<li><p>[SINGA-354] Update travis scripts to use conda-build for all platforms</p></li> +<li><p>[SINGA-358] Consolidated RUN steps and cleaned caches in Docker containers</p></li> +<li><p>[SINGA-359] Create alias for conda packages</p></li> </ul> </li> -<li>Documentation and usability<ul> -<li>[SINGA-223] Fix side navigation menu in the website</li> -<li>[SINGA-294] Add instructions to run CUDA unit tests on Windows</li> -<li>[SINGA-305] Add jupyter notebooks for SINGA V1 tutorial</li> -<li>[SINGA-319] Fix link errors on the index page</li> -<li>[SINGA-352] Complete SINGA documentation in Chinese version</li> -<li>[SINGA-361] Add git instructions for contributors and committers</li> +<li><p>Documentation and usability</p> +<ul> +<li><p>[SINGA-223] Fix side navigation menu in the website</p></li> +<li><p>[SINGA-294] Add instructions to run CUDA unit tests on Windows</p></li> +<li><p>[SINGA-305] Add jupyter notebooks for SINGA V1 tutorial</p></li> +<li><p>[SINGA-319] Fix link errors on the index page</p></li> +<li><p>[SINGA-352] Complete SINGA documentation in Chinese version</p></li> +<li><p>[SINGA-361] Add git instructions for contributors and committers</p></li> </ul> </li> -<li>Bugs fixed<ul> -<li>[SINGA-330] fix openblas building on i7 7700k</li> -<li>[SINGA-331] Fix the bug of tensor division operation</li> -<li>[SINGA-350] Error from python3 test</li> -<li>[SINGA-356] Error using travis tool to build SINGA on mac os</li> -<li>[SINGA-363] Fix some bugs in imagenet examples</li> -<li>[SINGA-368] Fix the bug in Cifar10 examples</li> -<li>[SINGA-369] the errors of examples in testing</li> +<li><p>Bugs fixed</p> +<ul> +<li><p>[SINGA-330] fix openblas building on i7 7700k</p></li> +<li><p>[SINGA-331] Fix the bug of tensor division operation</p></li> +<li><p>[SINGA-350] Error from python3 test</p></li> +<li><p>[SINGA-356] Error using travis tool to build SINGA on mac os</p></li> +<li><p>[SINGA-363] Fix some bugs in imagenet examples</p></li> +<li><p>[SINGA-368] Fix the bug in Cifar10 examples</p></li> +<li><p>[SINGA-369] the errors of examples in testing</p></li> </ul> </li> </ul> Modified: incubator/singa/site/trunk/releases/RELEASE_NOTES_2.0.0.html URL: http://svn.apache.org/viewvc/incubator/singa/site/trunk/releases/RELEASE_NOTES_2.0.0.html?rev=1862313&r1=1862312&r2=1862313&view=diff ============================================================================== --- incubator/singa/site/trunk/releases/RELEASE_NOTES_2.0.0.html (original) +++ incubator/singa/site/trunk/releases/RELEASE_NOTES_2.0.0.html Sat Jun 29 14:42:24 2019 @@ -204,45 +204,50 @@ learning models over large datasets.</p> <p>This release includes following features:</p> <ul class="simple"> -<li>Core components<ul> -<li>[SINGA-434] Support tensor broadcasting</li> -<li>[SINGA-370] Improvement to tensor reshape and various misc. changes related to SINGA-341 and 351</li> +<li><p>Core components</p> +<ul> +<li><p>[SINGA-434] Support tensor broadcasting</p></li> +<li><p>[SINGA-370] Improvement to tensor reshape and various misc. changes related to SINGA-341 and 351</p></li> </ul> </li> -<li>Model components<ul> -<li>[SINGA-333] Add support for Open Neural Network Exchange (ONNX) format</li> -<li>[SINGA-385] Add new python module for optimizers</li> -<li>[SINGA-394] Improve the CPP operations via Intel MKL DNN lib</li> -<li>[SINGA-425] Add 3 operators , Abs(), Exp() and leakyrelu(), for Autograd</li> -<li>[SINGA-410] Add two function, set_params() and get_params(), for Autograd Layer class</li> -<li>[SINGA-383] Add Separable Convolution for autograd</li> -<li>[SINGA-388] Develop some RNN layers by calling tiny operations like matmul, addbias.</li> -<li>[SINGA-382] Implement concat operation for autograd</li> -<li>[SINGA-378] Implement maxpooling operation and its related functions for autograd</li> -<li>[SINGA-379] Implement batchnorm operation and its related functions for autograd</li> +<li><p>Model components</p> +<ul> +<li><p>[SINGA-333] Add support for Open Neural Network Exchange (ONNX) format</p></li> +<li><p>[SINGA-385] Add new python module for optimizers</p></li> +<li><p>[SINGA-394] Improve the CPP operations via Intel MKL DNN lib</p></li> +<li><p>[SINGA-425] Add 3 operators , Abs(), Exp() and leakyrelu(), for Autograd</p></li> +<li><p>[SINGA-410] Add two function, set_params() and get_params(), for Autograd Layer class</p></li> +<li><p>[SINGA-383] Add Separable Convolution for autograd</p></li> +<li><p>[SINGA-388] Develop some RNN layers by calling tiny operations like matmul, addbias.</p></li> +<li><p>[SINGA-382] Implement concat operation for autograd</p></li> +<li><p>[SINGA-378] Implement maxpooling operation and its related functions for autograd</p></li> +<li><p>[SINGA-379] Implement batchnorm operation and its related functions for autograd</p></li> </ul> </li> -<li>Utility functions and CI<ul> -<li>[SINGA-432] Update depdent lib versions in conda-build config</li> -<li>[SINGA-429] Update docker images for latest cuda and cudnn</li> -<li>[SINGA-428] Move Docker images under Apache user name</li> +<li><p>Utility functions and CI</p> +<ul> +<li><p>[SINGA-432] Update depdent lib versions in conda-build config</p></li> +<li><p>[SINGA-429] Update docker images for latest cuda and cudnn</p></li> +<li><p>[SINGA-428] Move Docker images under Apache user name</p></li> </ul> </li> -<li>Documentation and usability<ul> -<li>[SINGA-395] Add documentation for autograd APIs</li> -<li>[SINGA-344] Add a GAN example</li> -<li>[SINGA-390] Update installation.md</li> -<li>[SINGA-384] Implement ResNet using autograd API</li> -<li>[SINGA-352] Complete SINGA documentation in Chinese version</li> +<li><p>Documentation and usability</p> +<ul> +<li><p>[SINGA-395] Add documentation for autograd APIs</p></li> +<li><p>[SINGA-344] Add a GAN example</p></li> +<li><p>[SINGA-390] Update installation.md</p></li> +<li><p>[SINGA-384] Implement ResNet using autograd API</p></li> +<li><p>[SINGA-352] Complete SINGA documentation in Chinese version</p></li> </ul> </li> -<li>Bugs fixed<ul> -<li>[SINGA-431] Unit Test failed - Tensor Transpose</li> -<li>[SINGA-422] ModuleNotFoundError: No module named â_singa_wrapâ</li> -<li>[SINGA-418] Unsupportive type âlongâ in python3.</li> -<li>[SINGA-409] Basic <code class="docutils literal notranslate"><span class="pre">singa-cpu</span></code> import throws error</li> -<li>[SINGA-408] Unsupportive function definition in python3</li> -<li>[SINGA-380] Fix bugs from Reshape</li> +<li><p>Bugs fixed</p> +<ul> +<li><p>[SINGA-431] Unit Test failed - Tensor Transpose</p></li> +<li><p>[SINGA-422] ModuleNotFoundError: No module named â_singa_wrapâ</p></li> +<li><p>[SINGA-418] Unsupportive type âlongâ in python3.</p></li> +<li><p>[SINGA-409] Basic <code class="docutils literal notranslate"><span class="pre">singa-cpu</span></code> import throws error</p></li> +<li><p>[SINGA-408] Unsupportive function definition in python3</p></li> +<li><p>[SINGA-380] Fix bugs from Reshape</p></li> </ul> </li> </ul>
