index.html 24.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
  <meta charset="utf-8">
  <meta http-equiv="X-UA-Compatible" content="IE=edge">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  
  
  <link rel="shortcut icon" href="../img/favicon.ico">
  <title>模型库2 - PaddleSlim Docs</title>
  <link href='https://fonts.googleapis.com/css?family=Lato:400,700|Roboto+Slab:400,700|Inconsolata:400,700' rel='stylesheet' type='text/css'>

  <link rel="stylesheet" href="../css/theme.css" type="text/css" />
  <link rel="stylesheet" href="../css/theme_extra.css" type="text/css" />
  <link rel="stylesheet" href="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/github.min.css">
  
  <script>
    // Current page data
    var mkdocs_page_name = "\u6a21\u578b\u5e932";
    var mkdocs_page_input_path = "model_zoo2.md";
    var mkdocs_page_url = null;
  </script>
  
  <script src="../js/jquery-2.1.1.min.js" defer></script>
  <script src="../js/modernizr-2.8.3.min.js" defer></script>
  <script src="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/highlight.min.js"></script>
  <script>hljs.initHighlightingOnLoad();</script> 
  
</head>

<body class="wy-body-for-nav" role="document">

  <div class="wy-grid-for-nav">

    
    <nav data-toggle="wy-nav-shift" class="wy-nav-side stickynav">
      <div class="wy-side-nav-search">
        <a href=".." class="icon icon-home"> PaddleSlim Docs</a>
        <div role="search">
  <form id ="rtd-search-form" class="wy-form" action="../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" title="Type search term here" />
  </form>
</div>
      </div>

      <div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
	<ul class="current">
	  
          
            <li class="toctree-l1">
		
    <a class="" href="..">Home</a>
	    </li>
          
            <li class="toctree-l1">
		
    <span class="caption-text">API</span>
    <ul class="subnav">
                <li class="">
                    
    <a class="" href="../api/quantization_api/">量化</a>
                </li>
                <li class="">
                    
    <a class="" href="../api/prune_api/">剪枝与敏感度</a>
                </li>
                <li class="">
                    
    <a class="" href="../api/analysis_api/">模型分析</a>
                </li>
                <li class="">
                    
    <a class="" href="../api/single_distiller_api/">知识蒸馏</a>
                </li>
                <li class="">
                    
    <a class="" href="../api/nas_api/">SA搜索</a>
                </li>
                <li class="">
                    
    <a class="" href="../api/search_space/">搜索空间</a>
                </li>
                <li class="">
                    
    <a class="" href="../table_latency/">硬件延时评估表</a>
                </li>
    </ul>
	    </li>
          
            <li class="toctree-l1">
		
    <span class="caption-text">教程</span>
    <ul class="subnav">
                <li class="">
                    
    <a class="" href="../tutorials/quant_post_demo/">离线量化</a>
                </li>
                <li class="">
                    
    <a class="" href="../tutorials/quant_aware_demo/">量化训练</a>
                </li>
                <li class="">
                    
    <a class="" href="../tutorials/quant_embedding_demo/">Embedding量化</a>
                </li>
                <li class="">
                    
    <a class="" href="../tutorials/nas_demo/">SA搜索</a>
                </li>
                <li class="">
                    
    <a class="" href="../tutorials/distillation_demo/">知识蒸馏</a>
                </li>
    </ul>
	    </li>
          
            <li class="toctree-l1">
		
    <a class="" href="../algo/algo/">算法原理</a>
	    </li>
          
            <li class="toctree-l1">
		
    <a class="" href="../model_zoo/">模型库</a>
	    </li>
          
            <li class="toctree-l1 current">
		
    <a class="current" href="./">模型库2</a>
    <ul class="subnav">
            
    <li class="toctree-l2"><a href="#1">1. 量化</a></li>
    
        <ul>
        
            <li><a class="toctree-l3" href="#11">1.1 图象分类</a></li>
        
139
            <li><a class="toctree-l3" href="#12">1.2 目标检测</a></li>
140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
        
            <li><a class="toctree-l3" href="#13">1.3 图像分割</a></li>
        
        </ul>
    

    <li class="toctree-l2"><a href="#2">2. 剪枝</a></li>
    
        <ul>
        
            <li><a class="toctree-l3" href="#21">2.1 图像分类</a></li>
        
            <li><a class="toctree-l3" href="#22">2.2 目标检测</a></li>
        
            <li><a class="toctree-l3" href="#23">2.3 图像分割</a></li>
        
        </ul>
    

    <li class="toctree-l2"><a href="#3">3. 蒸馏</a></li>
    
        <ul>
        
            <li><a class="toctree-l3" href="#31">3.1 图象分类</a></li>
        
            <li><a class="toctree-l3" href="#32">3.2 目标检测</a></li>
        
        </ul>
    

    </ul>
	    </li>
          
        </ul>
      </div>
      &nbsp;
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">

      
      <nav class="wy-nav-top" role="navigation" aria-label="top navigation">
        <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
        <a href="..">PaddleSlim Docs</a>
      </nav>

      
      <div class="wy-nav-content">
        <div class="rst-content">
          <div role="navigation" aria-label="breadcrumbs navigation">
  <ul class="wy-breadcrumbs">
    <li><a href="..">Docs</a> &raquo;</li>
    
      
    
    <li>模型库2</li>
    <li class="wy-breadcrumbs-aside">
      
        <a href="https://github.com/PaddlePaddle/PaddleSlim/edit/master/docs/model_zoo2.md"
          class="icon icon-github"> Edit on GitHub</a>
      
    </li>
  </ul>
  <hr/>
</div>
          <div role="main">
            <div class="section">
              
                <h2 id="1">1. 量化<a class="headerlink" href="#1" title="Permanent link">#</a></h2>
<h3 id="11">1.1 图象分类<a class="headerlink" href="#11" title="Permanent link">#</a></h3>
<p>数据集:ImageNet1000类</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">Top-1/Top-5</th>
<th align="center">模型大小(MB)</th>
217
<th align="center">下载</th>
218 219 220 221 222
</tr>
</thead>
<tbody>
<tr>
<td align="center">MobileNetV1 FP32</td>
223
<td align="center">70.99%/89.68%</td>
224
<td align="center">xx</td>
225
<td align="center"><a href="">下载链接</a></td>
226 227 228
</tr>
<tr>
<td align="center">MobileNetV1 quant_post</td>
229 230 231
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
232 233 234
</tr>
<tr>
<td align="center">MobileNetV1 quant_aware</td>
235 236 237
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
238 239 240
</tr>
<tr>
<td align="center">MobileNetV2 FP32</td>
241
<td align="center">72.15%/90.65%</td>
242
<td align="center">xx</td>
243
<td align="center"><a href="">下载链接</a></td>
244 245 246
</tr>
<tr>
<td align="center">MobileNetV2 quant_post</td>
247 248 249
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
250 251 252
</tr>
<tr>
<td align="center">MobileNetV2 quant_aware</td>
253 254 255
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
256 257 258
</tr>
<tr>
<td align="center">ResNet50 FP32</td>
259
<td align="center">76.50%/93.00%</td>
260
<td align="center">xx</td>
261
<td align="center"><a href="">下载链接</a></td>
262 263 264
</tr>
<tr>
<td align="center">ResNet50 quant_post</td>
265 266 267
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
268 269 270
</tr>
<tr>
<td align="center">ResNet50 quant_aware</td>
271 272 273
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
274 275 276
</tr>
</tbody>
</table>
277
<h3 id="12">1.2 目标检测<a class="headerlink" href="#12" title="Permanent link">#</a></h3>
278 279 280 281 282 283
<p>数据集:COCO 2017 </p>
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">Image/GPU</th>
284 285 286
<th align="center">输入608 Box AP</th>
<th align="center">输入416 Box AP</th>
<th align="center">输入320 Box AP</th>
287
<th align="center">模型大小(MB)</th>
288
<th align="center">下载</th>
289 290 291 292 293 294
</tr>
</thead>
<tbody>
<tr>
<td align="center">MobileNet-V1-YOLOv3 FP32</td>
<td align="center">8</td>
295 296 297
<td align="center">29.3</td>
<td align="center">29.3</td>
<td align="center">27.1</td>
298
<td align="center">xx</td>
299
<td align="center"><a href="">下载链接</a></td>
300 301 302
</tr>
<tr>
<td align="center">MobileNet-V1-YOLOv3 quant_post</td>
303 304 305 306 307 308
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
309 310 311
</tr>
<tr>
<td align="center">MobileNet-V1-YOLOv3 quant_aware</td>
312 313 314 315 316 317
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
318 319 320
</tr>
<tr>
<td align="center">R50-dcn-YOLOv3 FP32</td>
321 322 323 324 325 326
<td align="center">8</td>
<td align="center">41.4</td>
<td align="center">-</td>
<td align="center">-</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
327 328 329
</tr>
<tr>
<td align="center">R50-dcn-YOLOv3 quant_post</td>
330 331 332 333 334 335
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">-</td>
<td align="center">-</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
336 337 338
</tr>
<tr>
<td align="center">R50-dcn-YOLOv3 quant_aware</td>
339 340 341 342 343 344
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">-</td>
<td align="center">-</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
345 346 347
</tr>
</tbody>
</table>
348
<p>数据集:WIDER-FACE</p>
349 350 351 352 353
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">Image/GPU</th>
354 355 356 357
<th align="center">输入尺寸</th>
<th align="center">Easy/Medium/Hard</th>
<th align="center">模型大小(MB)</th>
<th align="center">下载</th>
358 359 360 361
</tr>
</thead>
<tbody>
<tr>
362
<td align="center">BlazeFace FP32</td>
363
<td align="center">8</td>
364 365
<td align="center">640</td>
<td align="center">0.915/0.892/0.797</td>
366
<td align="center">xx</td>
367
<td align="center"><a href="">下载链接</a></td>
368 369
</tr>
<tr>
370 371 372 373 374 375
<td align="center">BlazeFace quant_post</td>
<td align="center">8</td>
<td align="center">640</td>
<td align="center">xx/xx/xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
376 377
</tr>
<tr>
378
<td align="center">BlazeFace quant_aware</td>
379
<td align="center">8</td>
380 381 382 383
<td align="center">640</td>
<td align="center">xx/xx/xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
384 385
</tr>
<tr>
386 387
<td align="center">BlazeFace-Lite FP32</td>
<td align="center">8</td>
388
<td align="center">640</td>
389 390 391
<td align="center">0.909/0.885/0.781</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
392 393
</tr>
<tr>
394 395
<td align="center">BlazeFace-Lite quant_post</td>
<td align="center">8</td>
396
<td align="center">640</td>
397 398 399
<td align="center">xx/xx/xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
400 401
</tr>
<tr>
402 403 404 405 406 407
<td align="center">BlazeFace-Lite quant_aware</td>
<td align="center">8</td>
<td align="center">640</td>
<td align="center">xx/xx/xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
408 409
</tr>
<tr>
410 411 412 413 414 415
<td align="center">BlazeFace-NAS FP32</td>
<td align="center">8</td>
<td align="center">640</td>
<td align="center">0.837/0.807/0.658</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
416 417
</tr>
<tr>
418 419 420 421 422 423
<td align="center">BlazeFace-NAS quant_post</td>
<td align="center">8</td>
<td align="center">640</td>
<td align="center">xx/xx/xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
424 425
</tr>
<tr>
426 427 428 429 430 431
<td align="center">BlazeFace-NAS quant_aware</td>
<td align="center">8</td>
<td align="center">640</td>
<td align="center">xx/xx/xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
432 433 434
</tr>
</tbody>
</table>
435
<h3 id="_1"><a class="headerlink" href="#_1" title="Permanent link">#</a></h3>
436 437 438 439 440 441
<h3 id="13">1.3 图像分割<a class="headerlink" href="#13" title="Permanent link">#</a></h3>
<p>数据集:Cityscapes</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
442 443 444
<th align="center">mIoU</th>
<th align="center">模型大小(MB)</th>
<th align="center">下载</th>
445 446 447 448 449
</tr>
</thead>
<tbody>
<tr>
<td align="center">DeepLabv3+/MobileNetv1</td>
450 451 452
<td align="center">63.26</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
453 454
</tr>
<tr>
455 456 457 458
<td align="center">DeepLabv3+/MobileNetv1 quant_post</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
459 460
</tr>
<tr>
461 462 463 464
<td align="center">DeepLabv3+/MobileNetv1 quant_aware</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
465 466
</tr>
<tr>
467 468 469 470
<td align="center">DeepLabv3+/MobileNetv2</td>
<td align="center">69.81</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
471 472
</tr>
<tr>
473 474 475 476 477 478 479 480 481 482
<td align="center">DeepLabv3+/MobileNetv2 quant_post</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
</tr>
<tr>
<td align="center">DeepLabv3+/MobileNetv2 quant_aware</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
483 484 485
</tr>
</tbody>
</table>
486
<h3 id="_2"><a class="headerlink" href="#_2" title="Permanent link">#</a></h3>
487 488 489 490 491 492 493 494
<h2 id="2">2. 剪枝<a class="headerlink" href="#2" title="Permanent link">#</a></h2>
<h3 id="21">2.1 图像分类<a class="headerlink" href="#21" title="Permanent link">#</a></h3>
<p>数据集:ImageNet1000类</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">Top-1/Top-5</th>
495 496 497
<th align="center">模型大小(MB)</th>
<th align="center">FLOPs</th>
<th align="center">下载</th>
498 499 500 501 502
</tr>
</thead>
<tbody>
<tr>
<td align="center">MobileNetV1</td>
503 504 505 506
<td align="center">70.99%/89.68%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
507 508 509
</tr>
<tr>
<td align="center">MobileNetV1 uniform -50%</td>
510 511 512 513
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
514 515 516
</tr>
<tr>
<td align="center">MobileNetV1 sensitive -xx%</td>
517 518 519 520
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
521 522 523
</tr>
<tr>
<td align="center">MobileNetV2</td>
524 525 526 527
<td align="center">72.15%/90.65%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
528 529 530
</tr>
<tr>
<td align="center">MobileNetV2 uniform -50%</td>
531 532 533 534
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
535 536 537
</tr>
<tr>
<td align="center">MobileNetV2 sensitive -xx%</td>
538 539 540 541
<td align="center">xx%/xx%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
542 543 544
</tr>
<tr>
<td align="center">ResNet34</td>
545 546 547 548
<td align="center">74.57%/92.14%</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
549 550 551
</tr>
<tr>
<td align="center">ResNet34 uniform -50%</td>
552
<td align="center">xx%/xx%</td>
553 554
<td align="center">xx</td>
<td align="center">xx</td>
555
<td align="center"><a href="">下载链接</a></td>
556 557
</tr>
<tr>
558 559
<td align="center">ResNet34 auto -50%</td>
<td align="center">xx%/xx%</td>
560 561
<td align="center">xx</td>
<td align="center">xx</td>
562
<td align="center"><a href="">下载链接</a></td>
563 564 565
</tr>
</tbody>
</table>
566
<h3 id="_3"><a class="headerlink" href="#_3" title="Permanent link">#</a></h3>
567 568 569 570 571 572 573 574
<h3 id="22">2.2 目标检测<a class="headerlink" href="#22" title="Permanent link">#</a></h3>
<p>数据集:Pasacl VOC &amp; COCO 2017</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">数据集</th>
<th align="center">Image/GPU</th>
575 576 577 578 579 580
<th align="center">输入608 mAP</th>
<th align="center">输入416 mAP</th>
<th align="center">输入320  mAP</th>
<th align="center">模型大小(MB)</th>
<th align="center">FLOPs</th>
<th align="center">下载</th>
581 582 583 584 585 586 587
</tr>
</thead>
<tbody>
<tr>
<td align="center">MobileNet-V1-YOLOv3</td>
<td align="center">Pasacl VOC</td>
<td align="center">8</td>
588 589 590 591 592 593
<td align="center">76.2</td>
<td align="center">76.7</td>
<td align="center">75.3</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
594 595
</tr>
<tr>
596
<td align="center">MobileNet-V1-YOLOv3 prune xx%</td>
597
<td align="center">Pasacl VOC</td>
598 599 600 601 602 603 604
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
605 606 607 608
</tr>
<tr>
<td align="center">MobileNet-V1-YOLOv3</td>
<td align="center">COCO</td>
609 610 611 612 613 614 615
<td align="center">8</td>
<td align="center">29.3</td>
<td align="center">29.3</td>
<td align="center">27.1</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
616 617
</tr>
<tr>
618
<td align="center">MobileNet-V1-YOLOv3 prune xx%</td>
619
<td align="center">COCO</td>
620 621 622 623 624 625 626
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
627 628 629 630
</tr>
<tr>
<td align="center">R50-dcn-YOLOv3</td>
<td align="center">COCO</td>
631 632 633 634
<td align="center">8</td>
<td align="center">41.4</td>
<td align="center">-</td>
<td align="center">-</td>
635 636
<td align="center">xx</td>
<td align="center">xx</td>
637
<td align="center"><a href="">下载链接</a></td>
638 639
</tr>
<tr>
640 641 642 643 644 645
<td align="center">R50-dcn-YOLOv3 prune xx%</td>
<td align="center">COCO</td>
<td align="center">8</td>
<td align="center">xx</td>
<td align="center">-</td>
<td align="center">-</td>
646 647
<td align="center">xx</td>
<td align="center">xx</td>
648
<td align="center"><a href="">下载链接</a></td>
649 650 651 652 653 654 655 656 657
</tr>
</tbody>
</table>
<h3 id="23">2.3 图像分割<a class="headerlink" href="#23" title="Permanent link">#</a></h3>
<p>数据集:Cityscapes</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
658 659 660 661
<th align="center">mIoU</th>
<th align="center">模型大小(MB)</th>
<th align="center">FLOPs</th>
<th align="center">下载</th>
662 663 664 665 666
</tr>
</thead>
<tbody>
<tr>
<td align="center">DeepLabv3+/MobileNetv2</td>
667 668 669 670
<td align="center">69.81</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
671 672
</tr>
<tr>
673 674
<td align="center">DeepLabv3+/MobileNetv2 prune xx%</td>
<td align="center">xx</td>
675 676
<td align="center">xx</td>
<td align="center">xx</td>
677
<td align="center"><a href="">下载链接</a></td>
678 679 680 681 682 683 684 685 686 687 688
</tr>
</tbody>
</table>
<h2 id="3">3. 蒸馏<a class="headerlink" href="#3" title="Permanent link">#</a></h2>
<h3 id="31">3.1 图象分类<a class="headerlink" href="#31" title="Permanent link">#</a></h3>
<p>数据集:ImageNet1000类</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">baseline</th>
689
<th align="center">下载</th>
690 691 692 693 694
</tr>
</thead>
<tbody>
<tr>
<td align="center">MobileNetV1</td>
695 696 697 698 699 700 701
<td align="center">70.99%/89.68%</td>
<td align="center"><a href="http://paddle-imagenet-models-name.bj.bcebos.com/MobileNetV1_pretrained.tar">下载链接</a></td>
</tr>
<tr>
<td align="center">MobileNetV1 distilled (teacher: ResNet50_vd<sup><a href="#trans1">1</a></sup>)</td>
<td align="center">72.79%/90.69%</td>
<td align="center"><a href="">下载链接</a></td>
702 703 704
</tr>
<tr>
<td align="center">MobileNetV2</td>
705 706 707 708 709 710 711
<td align="center">72.15%/90.65%</td>
<td align="center"><a href="">下载链接</a></td>
</tr>
<tr>
<td align="center">MobileNetV2 distilled (teacher: ResNet50_vd)</td>
<td align="center">74.30%/91.52%</td>
<td align="center"><a href="">下载链接</a></td>
712 713 714
</tr>
<tr>
<td align="center">ResNet50</td>
715 716 717 718 719 720 721
<td align="center">76.50%/93.00%</td>
<td align="center"><a href="">下载链接</a></td>
</tr>
<tr>
<td align="center">ResNet50 distilled (teacher: ResNet101<sup><a href="#trans2">2</a></sup>)</td>
<td align="center">77.40%/93.48%</td>
<td align="center"><a href="">下载链接</a></td>
722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738
</tr>
</tbody>
</table>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p><a name="trans1">  [1]</a><a href="https://paddle-imagenet-models-name.bj.bcebos.com/ResNet50_vd_pretrained.tar">ResNet50_vd</a>预训练模型Top-1/Top-5准确率分别为79.12%/94.44%</p>
<p>带_vd后缀代表开启了Mixup训练,Mixup相关介绍参考<a href="https://arxiv.org/abs/1710.09412">mixup: Beyond Empirical Risk Minimization</a></p>
<p><a name="trans1">[2]</a><a href="https://paddle-imagenet-models-name.bj.bcebos.com/ResNet101_pretrained.tar">ResNet101</a>预训练模型Top-1/Top-5准确率分别为77.56%/93.64%</p>
</div>
<h3 id="32">3.2 目标检测<a class="headerlink" href="#32" title="Permanent link">#</a></h3>
<p>数据集:Pasacl VOC &amp; COCO 2017</p>
<table>
<thead>
<tr>
<th align="center">Model</th>
<th align="center">数据集</th>
<th align="center">Image/GPU</th>
739 740 741 742
<th align="center">输入640 mAP</th>
<th align="center">输入416 mAP</th>
<th align="center">输入320 mAP</th>
<th align="center">下载链接</th>
743 744 745 746 747 748 749
</tr>
</thead>
<tbody>
<tr>
<td align="center">MobileNet-V1-YOLOv3</td>
<td align="center">Pasacl VOC</td>
<td align="center">16</td>
750 751 752 753
<td align="center">76.2</td>
<td align="center">76.7</td>
<td align="center">75.3</td>
<td align="center"><a href="">下载链接</a></td>
754 755
</tr>
<tr>
756
<td align="center">MobileNet-V1-YOLOv3 distilled (teacher: ResNet34-YOLOv3-VOC<sup><a href="#trans3">3</a></sup>)</td>
757
<td align="center">Pasacl VOC</td>
758 759 760 761 762
<td align="center">16</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
763 764 765 766
</tr>
<tr>
<td align="center">MobileNet-V1-YOLOv3</td>
<td align="center">COCO</td>
767 768 769 770 771
<td align="center">16</td>
<td align="center">29.3</td>
<td align="center">29.3</td>
<td align="center">27.1</td>
<td align="center"><a href="">下载链接</a></td>
772 773
</tr>
<tr>
774
<td align="center">MobileNet-V1-YOLOv3 distilled (teacher: ResNet34-YOLOv3-COCO<sup><a href="#trans4">4</a></sup>)</td>
775
<td align="center">COCO</td>
776 777 778 779 780
<td align="center">16</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center">xx</td>
<td align="center"><a href="">下载链接</a></td>
781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837
</tr>
</tbody>
</table>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p><a name="trans1">[3]</a><a href="">ResNet34-YOLOv3-VOC</a>预训练模型的Box AP为82.6</p>
<p><a name="trans1">[4]</a><a href="">ResNet34-YOLOv3-COCO</a>预训练模型的Box AP为36.2</p>
</div>
              
            </div>
          </div>
          <footer>
  
    <div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
      
      
        <a href="../model_zoo/" class="btn btn-neutral" title="模型库"><span class="icon icon-circle-arrow-left"></span> Previous</a>
      
    </div>
  

  <hr/>

  <div role="contentinfo">
    <!-- Copyright etc -->
    
  </div>

  Built with <a href="http://www.mkdocs.org">MkDocs</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
      
        </div>
      </div>

    </section>

  </div>

  <div class="rst-versions" role="note" style="cursor: pointer">
    <span class="rst-current-version" data-toggle="rst-current-version">
      
          <a href="https://github.com/PaddlePaddle/PaddleSlim/" class="fa fa-github" style="float: left; color: #fcfcfc"> GitHub</a>
      
      
        <span><a href="../model_zoo/" style="color: #fcfcfc;">&laquo; Previous</a></span>
      
      
    </span>
</div>
    <script>var base_url = '..';</script>
    <script src="../js/theme.js" defer></script>
      <script src="../mathjax-config.js" defer></script>
      <script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML" defer></script>
      <script src="../search/main.js" defer></script>

</body>
</html>