mirrors/deep-person-reid
1
0
Fork 0
mirror of https://github.com/KaiyangZhou/deep-person-reid.git synced 2025-06-03 14:53:23 +08:00
Code Issues Projects Releases Wiki Activity

update model zoo

Browse Source
This commit is contained in:
KaiyangZhou 2019-05-28 15:33:20 +01:00
parent 6b0eb9a5f7
commit ebcaf80276
6 changed files with 68 additions and 105 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

121
MODEL_ZOO.html
View File

@ -170,17 +170,17 @@
<div class="section" id="model-zoo">
<h1>Model Zoo<a class="headerlink" href="#model-zoo" title="Permalink to this headline"></a></h1>
<ul class="simple">
<li>Results are presented in the format of <em>Rank-1 (mAP)</em>.</li>
<li>Classification layer is ignored when computing the model size.</li>
<li>Results are presented in the format of <em>&lt;Rank-1 (mAP)&gt;</em>.</li>
<li>When computing FLOPs, only layers that are used at test time are considered (see <code class="docutils literal notranslate"><span class="pre">torchreid.utils.compute_model_complexity</span></code>).</li>
<li>Unless specified otherwise, only <code class="docutils literal notranslate"><span class="pre">Random2DTranslation</span></code> and <code class="docutils literal notranslate"><span class="pre">RandomHorizontalFlip</span></code> are used for data augmentation.</li>
<li>Asterisk (*) means the model is trained from scratch.</li>
<li>Model weights can be downloaded by clicking the highlighted numbers.</li>
</ul>
<table border="1" class="docutils">
<thead>
<tr>
<th align="left">Model</th>
<th align="center"># param (M)</th>
<th align="center"># Param (10^6)</th>
<th align="center">GFLOPs</th>
<th align="center">Loss</th>
<th align="center">Input</th>
<th align="center">market1501</th>
@ -192,101 +192,62 @@
<tr>
<td align="left">resnet50</td>
<td align="center">23.5</td>
<td align="center">2.7</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_market_xent.zip">87.9 (70.4)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_duke_xent.zip">78.3 (58.9)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_msmt_xent.zip">63.2 (33.9)</a></td>
<td align="center"><a href="https://mega.nz/#!FKZjVKaZ!4v_FR8pTvuHoMQIKdstJ_YCsRrtZW2hwWxc-T0JIlHE">87.9 (70.4)</a></td>
<td align="center"><a href="https://mega.nz/#!JPZjCYhK!YVJbE_4vTc8DX19Rt_FB77YY4BaEA1P6Xb5sNJGep2M">78.3 (58.9)</a></td>
<td align="center"><a href="https://mega.nz/#!APAxDY4Z!Iou9x8s3ATdYS2SlK2oiJbHrhvlzH7F1gE2qjM-GJGw">63.2 (33.9)</a></td>
</tr>
<tr>
<td align="left">resnet50_fc512</td>
<td align="center">24.6</td>
<td align="center">4.1</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_fc512_market_xent.zip">90.8 (75.3)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_fc512_duke_xent.zip">81.0 (64.0)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_fc512_msmt_xent.zip">69.6 (38.4)</a></td>
</tr>
<tr>
<td align="left">densenet121_fc512</td>
<td align="center">7.5</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/densenet121_fc512_market_xent.zip">87.8 (68.0)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/densenet121_fc512_duke_xent.zip">79.7 (58.8)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/densenet121_fc512_msmt_xent.zip">67.6 (35.0)</a></td>
</tr>
<tr>
<td align="left">se_resnet50_fc512</td>
<td align="center">27.1</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/se_resnet50_fc512_market_xent.zip">91.9 (75.8)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/se_resnet50_fc512_duke_xent.zip">81.5 (63.7)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/se_resnet50_fc512_msmt_xent.zip">71.1 (39.8)</a></td>
</tr>
<tr>
<td align="left">shufflenet</td>
<td align="center">0.9</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/shufflenet_market_xent.zip">84.1(64.1)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/shufflenet_duke_xent.zip">73.4(51.9)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/shufflenet_msmt_xent.zip">51.3(24.2)</a></td>
</tr>
<tr>
<td align="left">squeezenet1_0_fc512</td>
<td align="center">1.0</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/squeezenet1_0_fc512_market_xent.zip">79.3 (52.2)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/squeezenet1_0_fc512_duke_xent.zip">66.6 (42.6)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/squeezenet1_0_fc512_msmt_xent.zip">44.1 (17.1)</a></td>
</tr>
<tr>
<td align="left">mobilenetv2_1dot0</td>
<td align="center">2.2</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot0_market.pth.tar">85.6 (67.3)</a></td>
<td align="center"><a href="http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot0_duke.pth.tar">74.2 (54.7)</a></td>
<td align="center"><a href="http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot0_msmt.pth.tar">57.4 (29.3)</a></td>
</tr>
<tr>
<td align="left">mobilenetv2_1dot4</td>
<td align="center">4.3</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot4_market.pth.tar">87.0 (68.5)</a></td>
<td align="center"><a href="http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot4_duke.pth.tar">76.2 (55.8)</a></td>
<td align="center"><a href="http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot4_msmt.pth.tar">60.1 (31.5)</a></td>
</tr>
<tr>
<td align="left">resnet50mid</td>
<td align="center">27.7</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50mid_market_xent.zip">90.2 (76.0)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50mid_duke_xent.zip">81.6 (64.0)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50mid_msmt_xent.zip">69.0 (38.0)</a></td>
<td align="center"><a href="https://mega.nz/#!EaZjhKyS!lBvD3vAJ4DOmElZkNa7gyPM1RE661GUd2v9kK84gSZE">90.8 (75.3)</a></td>
<td align="center"><a href="https://mega.nz/#!lXYDSKZa!lumiXkY2H5Sm8gEgTWPBdWKv3ujy4zjrffjERaXkc9I">81.0 (64.0)</a></td>
<td align="center"><a href="https://mega.nz/#!9PQTXIpL!iI5wgieTCn0Jm-pyg9RCu0RkH43pV3ntHhr1PeqSyT4">69.6 (38.4)</a></td>
</tr>
<tr>
<td align="left">mlfn</td>
<td align="center">32.5</td>
<td align="center">2.8</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mlfn_market_xent.zip">90.1 (74.3)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mlfn_duke_xent.zip">81.1 (63.2)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mlfn_msmt_xent.zip">66.4 (37.2)</a></td>
<td align="center"><a href="https://mega.nz/#!kHQ3ESLT!NoGc8eHEBZOJZM19THh3DFfRBXIPXzM-sdLmF1mvTXA">90.1 (74.3)</a></td>
<td align="center"><a href="https://mega.nz/#!8PQXUCaI!mJO1vD9tI739hkNBj2QWUt0VPcZ-s89fSMMGPPP1msc">81.1 (63.2)</a></td>
<td align="center"><a href="https://mega.nz/#!paIXFQCS!W3ZGkxyF1idwvQzTRDE2p0DhNDki2SBJRfp7S_Cwphk">66.4 (37.2)</a></td>
</tr>
<tr>
<td align="left">hacnn<sup>*</sup></td>
<td align="center">3.7</td>
<td align="center">4.5</td>
<td align="center">0.5</td>
<td align="center">softmax</td>
<td align="center">(160, 64)</td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/hacnn_market_xent.zip">90.9 (75.6)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/hacnn_duke_xent.zip">80.1 (63.2)</a></td>
<td align="center"><a href="http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/hacnn_msmt_xent.zip">64.7 (37.2)</a></td>
<td align="center"><a href="https://mega.nz/#!ULQXUQBK!S-8v_pR2xBD3ZpuY0I7Bqift-eX_V84gajHMDG6zUac">90.9 (75.6)</a></td>
<td align="center"><a href="https://mega.nz/#!wPJTkAQR!XkKd39lsmBZMrCh3JjF6vnNafBZkouVIVdeBqQKdSzA">80.1 (63.2)</a></td>
<td align="center"><a href="https://mega.nz/#!AXAziKjL!JtMwHz2UYy58gDMQLGakSmF3JOr72o8zmkqlQA-LIpQ">64.7 (37.2)</a></td>
</tr>
<tr>
<td align="left">mobilenetv2_1dot0</td>
<td align="center">2.2</td>
<td align="center">0.2</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="https://mega.nz/#!8KYTFAIB!3dL35WQLxSoTSClDTv0kxa81k3fh5hXmAWA4_a3qiOI">85.6 (67.3)</a></td>
<td align="center"><a href="https://mega.nz/#!hbRXDSCL!YYgqJ6PVUf4clgtUuK2s5FRhYJdU3yTibLscwOTNnDk">74.2 (54.7)</a></td>
<td align="center"><a href="https://mega.nz/#!5SJTmCYb!ZQ8O2MN9JF4-WDAeX04Xex1KyuBYQ_o2aoMIsTgQ748">57.4 (29.3)</a></td>
</tr>
<tr>
<td align="left">mobilenetv2_1dot4</td>
<td align="center">4.3</td>
<td align="center">0.4</td>
<td align="center">softmax</td>
<td align="center">(256, 128)</td>
<td align="center"><a href="https://mega.nz/#!4XZhEKCS!6lTuTRbHIWU5nzJzTPDGykA7sPME8_1ISGsUYFJXZWA">87.0 (68.5)</a></td>
<td align="center"><a href="https://mega.nz/#!JbQVDIYQ!-7pnjIfpIDt1EoQOvpvuIEcTj3Qg8SE6o_3ZPGWrIcw">76.2 (55.8)</a></td>
<td align="center"><a href="https://mega.nz/#!gOYDAQrK!sMJO7c_X4iIxoVfV_tXYdzeDJByPo5XkUjEN7Z2JTmM">60.1 (31.5)</a></td>
</tr>
</tbody>
</table></div>

6
_modules/torchreid/models/mlfn.html
View File

@ -179,7 +179,7 @@
<span class="n">model_urls</span> <span class="o">=</span> <span class="p">{</span>
<span class="c1"># training epoch = 5, top1 = 51.6</span>
<span class="s1">&#39;imagenet&#39;</span><span class="p">:</span> <span class="s1">&#39;http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/imagenet-pretrained/mlfn-9cb5a267.pth.tar&#39;</span><span class="p">,</span>
<span class="s1">&#39;imagenet&#39;</span><span class="p">:</span> <span class="s1">&#39;https://mega.nz/#!YHxAhaxC!yu9E6zWl0x5zscSouTdbZu8gdFFytDdl-RAdD2DEfpk&#39;</span><span class="p">,</span>
<span class="p">}</span>
@ -376,7 +376,9 @@
<span class="k">def</span> <span class="nf">mlfn</span><span class="p">(</span><span class="n">num_classes</span><span class="p">,</span> <span class="n">loss</span><span class="o">=</span><span class="s1">&#39;softmax&#39;</span><span class="p">,</span> <span class="n">pretrained</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">MLFN</span><span class="p">(</span><span class="n">num_classes</span><span class="p">,</span> <span class="n">loss</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="k">if</span> <span class="n">pretrained</span><span class="p">:</span>
<span class="n">init_pretrained_weights</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;imagenet&#39;</span><span class="p">])</span>
<span class="c1">#init_pretrained_weights(model, model_urls[&#39;imagenet&#39;])</span>
<span class="kn">import</span> <span class="nn">warnings</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="s1">&#39;The imagenet pretrained weights need to be manually downloaded from </span><span class="si">{}</span><span class="s1">&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;imagenet&#39;</span><span class="p">]))</span>
<span class="k">return</span> <span class="n">model</span>
</pre></div>

12
_modules/torchreid/models/mobilenetv2.html
View File

@ -178,9 +178,9 @@
<span class="n">model_urls</span> <span class="o">=</span> <span class="p">{</span>
<span class="c1"># 1.0: top-1 71.3</span>
<span class="s1">&#39;mobilenetv2_1dot0&#39;</span><span class="p">:</span> <span class="s1">&#39;http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/imagenet-pretrained/mobilenetv2_1.0-0f96a698.pth&#39;</span><span class="p">,</span>
<span class="s1">&#39;mobilenetv2_1dot0&#39;</span><span class="p">:</span> <span class="s1">&#39;https://mega.nz/#!NKp2wAIA!1NH1pbNzY_M2hVk_hdsxNM1NUOWvvGPHhaNr-fASF6c&#39;</span><span class="p">,</span>
<span class="c1"># 1.4: top-1 73.9</span>
<span class="s1">&#39;mobilenetv2_1dot4&#39;</span><span class="p">:</span> <span class="s1">&#39;http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/imagenet-pretrained/mobilenetv2_1.4-bc1cc36b.pth&#39;</span><span class="p">,</span>
<span class="s1">&#39;mobilenetv2_1dot4&#39;</span><span class="p">:</span> <span class="s1">&#39;https://mega.nz/#!RGhgEIwS!xN2s2ZdyqI6vQ3EwgmRXLEW3khr9tpXg96G9SUJugGk&#39;</span><span class="p">,</span>
<span class="p">}</span>
@ -377,7 +377,9 @@
<span class="o">**</span><span class="n">kwargs</span>
<span class="p">)</span>
<span class="k">if</span> <span class="n">pretrained</span><span class="p">:</span>
<span class="n">init_pretrained_weights</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;mobilenetv2_1dot0&#39;</span><span class="p">])</span>
<span class="c1">#init_pretrained_weights(model, model_urls[&#39;mobilenetv2_1dot0&#39;])</span>
<span class="kn">import</span> <span class="nn">warnings</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="s1">&#39;The imagenet pretrained weights need to be manually downloaded from </span><span class="si">{}</span><span class="s1">&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;mobilenetv2_1dot0&#39;</span><span class="p">]))</span>
<span class="k">return</span> <span class="n">model</span>
@ -391,7 +393,9 @@
<span class="o">**</span><span class="n">kwargs</span>
<span class="p">)</span>
<span class="k">if</span> <span class="n">pretrained</span><span class="p">:</span>
<span class="n">init_pretrained_weights</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;mobilenetv2_1dot4&#39;</span><span class="p">])</span>
<span class="c1">#init_pretrained_weights(model, model_urls[&#39;mobilenetv2_1dot4&#39;])</span>
<span class="kn">import</span> <span class="nn">warnings</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="s1">&#39;The imagenet pretrained weights need to be manually downloaded from </span><span class="si">{}</span><span class="s1">&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;mobilenetv2_1dot4&#39;</span><span class="p">]))</span>
<span class="k">return</span> <span class="n">model</span>
</pre></div>

6
_modules/torchreid/models/shufflenet.html
View File

@ -179,7 +179,7 @@
<span class="n">model_urls</span> <span class="o">=</span> <span class="p">{</span>
<span class="c1"># training epoch = 90, top1 = 61.8</span>
<span class="s1">&#39;imagenet&#39;</span><span class="p">:</span> <span class="s1">&#39;http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/imagenet-pretrained/shufflenet-bee1b265.pth.tar&#39;</span><span class="p">,</span>
<span class="s1">&#39;imagenet&#39;</span><span class="p">:</span> <span class="s1">&#39;https://mega.nz/#!RDpUlQCY!tr_5xBEkelzDjveIYBBcGcovNCOrgfiJO9kiidz9fZM&#39;</span><span class="p">,</span>
<span class="p">}</span>
@ -328,7 +328,9 @@
<span class="k">def</span> <span class="nf">shufflenet</span><span class="p">(</span><span class="n">num_classes</span><span class="p">,</span> <span class="n">loss</span><span class="o">=</span><span class="s1">&#39;softmax&#39;</span><span class="p">,</span> <span class="n">pretrained</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">):</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">ShuffleNet</span><span class="p">(</span><span class="n">num_classes</span><span class="p">,</span> <span class="n">loss</span><span class="p">,</span> <span class="o">**</span><span class="n">kwargs</span><span class="p">)</span>
<span class="k">if</span> <span class="n">pretrained</span><span class="p">:</span>
<span class="n">init_pretrained_weights</span><span class="p">(</span><span class="n">model</span><span class="p">,</span> <span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;imagenet&#39;</span><span class="p">])</span>
<span class="c1">#init_pretrained_weights(model, model_urls[&#39;imagenet&#39;])</span>
<span class="kn">import</span> <span class="nn">warnings</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="s1">&#39;The imagenet pretrained weights need to be manually downloaded from </span><span class="si">{}</span><span class="s1">&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">model_urls</span><span class="p">[</span><span class="s1">&#39;imagenet&#39;</span><span class="p">]))</span>
<span class="k">return</span> <span class="n">model</span>
</pre></div>

26
_sources/MODEL_ZOO.md.txt
View File

@ -1,21 +1,15 @@
# Model Zoo
- Results are presented in the format of *Rank-1 (mAP)*.
- Classification layer is ignored when computing the model size.
- Results are presented in the format of *<Rank-1 (mAP)>*.
- When computing FLOPs, only layers that are used at test time are considered (see `torchreid.utils.compute_model_complexity`).
- Unless specified otherwise, only `Random2DTranslation` and `RandomHorizontalFlip` are used for data augmentation.
- Asterisk (\*) means the model is trained from scratch.
- Model weights can be downloaded by clicking the highlighted numbers.
| Model | # param (M) | Loss | Input | market1501 | dukemtmcreid | msmt17 |
| :--- | :---: | :---: | :---: | :---: | :---: | :---: |
| resnet50 | 23.5 | softmax | (256, 128) | [87.9 (70.4)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_market_xent.zip) | [78.3 (58.9)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_duke_xent.zip) | [63.2 (33.9)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_msmt_xent.zip) |
| resnet50_fc512 | 24.6 | softmax | (256, 128) | [90.8 (75.3)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_fc512_market_xent.zip) | [81.0 (64.0)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_fc512_duke_xent.zip) | [69.6 (38.4)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50_fc512_msmt_xent.zip) |
| densenet121_fc512 | 7.5 | softmax | (256, 128) | [87.8 (68.0)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/densenet121_fc512_market_xent.zip) | [79.7 (58.8)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/densenet121_fc512_duke_xent.zip) | [67.6 (35.0)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/densenet121_fc512_msmt_xent.zip) |
| se_resnet50_fc512 | 27.1 | softmax | (256, 128) | [91.9 (75.8)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/se_resnet50_fc512_market_xent.zip) | [81.5 (63.7)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/se_resnet50_fc512_duke_xent.zip) | [71.1 (39.8)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/se_resnet50_fc512_msmt_xent.zip) |
| shufflenet | 0.9 | softmax | (256, 128) | [84.1(64.1)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/shufflenet_market_xent.zip) | [73.4(51.9)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/shufflenet_duke_xent.zip) | [51.3(24.2)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/shufflenet_msmt_xent.zip) |
| squeezenet1_0_fc512 | 1.0 | softmax | (256, 128) | [79.3 (52.2)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/squeezenet1_0_fc512_market_xent.zip) | [66.6 (42.6)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/squeezenet1_0_fc512_duke_xent.zip) | [44.1 (17.1)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/squeezenet1_0_fc512_msmt_xent.zip) |
| mobilenetv2_1dot0 | 2.2 | softmax | (256, 128) | [85.6 (67.3)](http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot0_market.pth.tar) | [74.2 (54.7)](http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot0_duke.pth.tar) | [57.4 (29.3)](http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot0_msmt.pth.tar) |
| mobilenetv2_1dot4 | 4.3 | softmax | (256, 128) | [87.0 (68.5)](http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot4_market.pth.tar) | [76.2 (55.8)](http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot4_duke.pth.tar) | [60.1 (31.5)](http://eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mobilenetv2_1dot4_msmt.pth.tar) |
| resnet50mid | 27.7 | softmax | (256, 128) | [90.2 (76.0)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50mid_market_xent.zip) | [81.6 (64.0)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50mid_duke_xent.zip) | [69.0 (38.0)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/resnet50mid_msmt_xent.zip) |
| mlfn | 32.5 | softmax | (256, 128) | [90.1 (74.3)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mlfn_market_xent.zip) | [81.1 (63.2)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mlfn_duke_xent.zip) | [66.4 (37.2)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/mlfn_msmt_xent.zip) |
| hacnn<sup>*</sup> | 3.7 | softmax | (160, 64) | [90.9 (75.6)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/hacnn_market_xent.zip) | [80.1 (63.2)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/hacnn_duke_xent.zip) | [64.7 (37.2)](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/model-zoo/image-models/hacnn_msmt_xent.zip) |
| Model | # Param (10^6) | GFLOPs | Loss | Input | market1501 | dukemtmcreid | msmt17 |
| :--- | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| resnet50 | 23.5 | 2.7 | softmax | (256, 128) | [87.9 (70.4)](https://mega.nz/#!FKZjVKaZ!4v_FR8pTvuHoMQIKdstJ_YCsRrtZW2hwWxc-T0JIlHE) | [78.3 (58.9)](https://mega.nz/#!JPZjCYhK!YVJbE_4vTc8DX19Rt_FB77YY4BaEA1P6Xb5sNJGep2M) | [63.2 (33.9)](https://mega.nz/#!APAxDY4Z!Iou9x8s3ATdYS2SlK2oiJbHrhvlzH7F1gE2qjM-GJGw) |
| resnet50_fc512 | 24.6 | 4.1 | softmax | (256, 128) | [90.8 (75.3)](https://mega.nz/#!EaZjhKyS!lBvD3vAJ4DOmElZkNa7gyPM1RE661GUd2v9kK84gSZE) | [81.0 (64.0)](https://mega.nz/#!lXYDSKZa!lumiXkY2H5Sm8gEgTWPBdWKv3ujy4zjrffjERaXkc9I) | [69.6 (38.4)](https://mega.nz/#!9PQTXIpL!iI5wgieTCn0Jm-pyg9RCu0RkH43pV3ntHhr1PeqSyT4) |
| mlfn | 32.5 | 2.8 | softmax | (256, 128) | [90.1 (74.3)](https://mega.nz/#!kHQ3ESLT!NoGc8eHEBZOJZM19THh3DFfRBXIPXzM-sdLmF1mvTXA) | [81.1 (63.2)](https://mega.nz/#!8PQXUCaI!mJO1vD9tI739hkNBj2QWUt0VPcZ-s89fSMMGPPP1msc) | [66.4 (37.2)](https://mega.nz/#!paIXFQCS!W3ZGkxyF1idwvQzTRDE2p0DhNDki2SBJRfp7S_Cwphk) |
| hacnn<sup>*</sup> | 4.5 | 0.5 | softmax | (160, 64) | [90.9 (75.6)](https://mega.nz/#!ULQXUQBK!S-8v_pR2xBD3ZpuY0I7Bqift-eX_V84gajHMDG6zUac) | [80.1 (63.2)](https://mega.nz/#!wPJTkAQR!XkKd39lsmBZMrCh3JjF6vnNafBZkouVIVdeBqQKdSzA) | [64.7 (37.2)](https://mega.nz/#!AXAziKjL!JtMwHz2UYy58gDMQLGakSmF3JOr72o8zmkqlQA-LIpQ) |
| mobilenetv2_1dot0 | 2.2 | 0.2 | softmax | (256, 128) | [85.6 (67.3)](https://mega.nz/#!8KYTFAIB!3dL35WQLxSoTSClDTv0kxa81k3fh5hXmAWA4_a3qiOI) | [74.2 (54.7)](https://mega.nz/#!hbRXDSCL!YYgqJ6PVUf4clgtUuK2s5FRhYJdU3yTibLscwOTNnDk) | [57.4 (29.3)](https://mega.nz/#!5SJTmCYb!ZQ8O2MN9JF4-WDAeX04Xex1KyuBYQ_o2aoMIsTgQ748) |
| mobilenetv2_1dot4 | 4.3 | 0.4 | softmax | (256, 128) | [87.0 (68.5)](https://mega.nz/#!4XZhEKCS!6lTuTRbHIWU5nzJzTPDGykA7sPME8_1ISGsUYFJXZWA) | [76.2 (55.8)](https://mega.nz/#!JbQVDIYQ!-7pnjIfpIDt1EoQOvpvuIEcTj3Qg8SE6o_3ZPGWrIcw) | [60.1 (31.5)](https://mega.nz/#!gOYDAQrK!sMJO7c_X4iIxoVfV_tXYdzeDJByPo5XkUjEN7Z2JTmM) |

2
searchindex.js
View File

File diff suppressed because one or more lines are too long