├── LICENSE
├── README.md
├── css
    └── style.css
├── images
    ├── aizoocom.png
    ├── demo.jpg
    └── result.png
├── index.html
├── js
    ├── detection.js
    ├── index.js
    └── tfjs.min.js
└── tfjs-models
    ├── group1-shard1of1.bin
    └── model.json


/LICENSE:
--------------------------------------------------------------------------------
 1 | MIT License
 2 | 
 3 | Copyright (c) 2020 AIZOOTech
 4 | 
 5 | Permission is hereby granted, free of charge, to any person obtaining a copy
 6 | of this software and associated documentation files (the "Software"), to deal
 7 | in the Software without restriction, including without limitation the rights
 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 9 | copies of the Software, and to permit persons to whom the Software is
10 | furnished to do so, subject to the following conditions:
11 | 
12 | The above copyright notice and this permission notice shall be included in all
13 | copies or substantial portions of the Software.
14 | 
15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21 | SOFTWARE.
22 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | # face-mask-detection-web-demo
 2 | # 人脸口罩检测网页demo
 3 | Face mask deteciton demo, by tensorflow.js
 4 | 
 5 | ## 原理
 6 | 本demo是在浏览器运行的人脸口罩检测网页demo，介绍如何将深度学习的人脸口罩检测模型部署到浏览器里面。
 7 | 关于人脸口罩检测的PyTorch、TensorFlow、Caffe、Keras、MXNet版本，可以进入相应Github仓库
 8 | [FaceMaskDetection](https://github.com/AIZOOTech/FaceMaskDetection)
 9 | 关于项目介绍，可以阅读一下两篇文章：
10 | [AIZOO开源人脸口罩检测数据+模型+代码+在线网页体验，通通都开源了](https://mp.weixin.qq.com/s/22U_v6IQ9PBHslI-65v_0Q)
11 | [人脸口罩检测现开源PyTorch、TensorFlow、MXNet等全部五大主流深度学习框架模型和代码
12 | ](https://mp.weixin.qq.com/s?__biz=MzIyMDY2MTUyNg==&mid=2247483779&idx=1&sn=b9ac5af31adf1dfdc3c87eb1c74836a5&exportkey=AX%2FANiIY8CWWMPQrHKh6A5E%3D&pass_ticket=aaNfWJGBgSum6CY5pvFqx0IIfljPPkeX%2BdMtPEl3zo5hQfPnYR5mEUlayz06kNKG)
13 | 
14 | 
15 | 深度学习模型，可以借助TensorFlow.js库，运行在浏览器里。首先，需要使用`tensorflowjs_converter`将tensorflow的`graph model`或者keras的`layer model`转换为TensorFlow.js支持的模型。
16 | 该工具可以通过`pip install tensorflowjs`安装。
17 | 
18 | 如果使用Keras模型转的操作如下：
19 | ```
20 | tensorflowjs_convert --input_format keras --output_format tfjs_layers_model   /path/to/keras/hdf5/model  /path/to/output/folder
21 | ```
22 | 模型会生成`model.json`文件和一个或者多个`bin`文件，其中前者保存模型的拓扑，后者保存模型的权重。
23 | 使用JavaScript，需要在html中先引入`tfjs.min.js`库，然后加载模型
24 | ```
25 | <script src="js/tfjs.min.js"></script> 
26 | 
27 | ```
28 | 在`detection.js`中，加载模型
29 | ```
30 | model = await tf.loadLayersModel('./tfjs-models/model.json');
31 | ```
32 | 置于anchor生成、输出解码、nms与使用python版本并无太大差异，大家可以查看`detection.js`中三个相关的函数，一目了然。
33 | 
34 | ## 运行方法
35 | 在当前目录下打开终端，只需要建立一个最小web server即可。
36 | 对于使用python的用户
37 | ```
38 | // python3用户
39 | python -m http.server
40 | // python2用户
41 | python -m SimpleHTTPServer
42 | 
43 | ```
44 | 如果你使用Node.js
45 | ```
46 | npm install serve -g //安装serve
47 | serve // this will open a mini web serve
48 | // 您也可以使用http-serve
49 | npm install http-server -g
50 | http-server
51 | ```
52 | 
53 | ## 效果
54 | 您可以点击网页中的上传图片按钮，或者拖拽图片到网页区域，然后模型会自动进行检测并画框。
55 | ![页面效果图](/images/result.png)


--------------------------------------------------------------------------------
/css/style.css:
--------------------------------------------------------------------------------
 1 | 
 2 | html, body {
 3 |   margin: 0;
 4 |   padding: 0;
 5 |   font-family: Apercu;
 6 |   
 7 | }
 8 | 
 9 | .title {
10 |   text-align: center
11 | }
12 | 
13 | #tips {
14 |   color:rosybrown
15 | }
16 | 
17 | #container {
18 |   background: #c79ef812;
19 |   /* height: 500px; */
20 |   padding: 50px;
21 |   font-size: 14px;
22 |   border-radius: 33px;
23 |   text-align: center;
24 | }
25 | 
26 | #text-container {
27 |   height: 100px;
28 |   width: 100%;
29 |   text-align: center;
30 | }
31 | 
32 | #instruction {
33 |   margin-top: 20px;
34 | }
35 | 
36 | .result-color {
37 |   color: #A159FC;
38 | }
39 | 
40 | .green-color {
41 |   color: #0bcf82;
42 | }
43 | 
44 | .highlight {
45 |   opacity: 0.4;
46 |   background-color: rgba(161,89,252, 0.2) !important;
47 | }
48 | 
49 | .center {
50 |   display: block;
51 |   margin-left: auto;
52 |   margin-right: auto;
53 | }
54 | 
55 | #fileUploader {
56 |   opacity: 0;
57 | }
58 | 
59 | #uploader-btn {
60 |   background: #faebd700;
61 |   margin-top: 1em;
62 |   color: #0bcf81;
63 |   border: solid 1px #0bcf81;
64 |   padding: .6em 2em;
65 |   line-height: 2;
66 |   font-size: 12px;
67 |   font-family: Apercu;
68 |   cursor: pointer;
69 |   border-radius: 10px;
70 | }
71 | 
72 | #uploader-btn:hover {
73 |   color: white;
74 |   background-color: #0bcf82;
75 |   border: solid 1px #0bcf82;
76 |   -webkit-transition: all 0.3s ease;
77 |   -moz-transition: all 0.3s ease;
78 |   -o-transition: all 0.3s ease;
79 |   transition: all 0.3s ease;
80 | }
81 | 


--------------------------------------------------------------------------------
/images/aizoocom.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/AIZOOTech/mask-detection-web-demo/b4425bc5fa5a75bd52e17a53a1a0b0243f8e5e7a/images/aizoocom.png


--------------------------------------------------------------------------------
/images/demo.jpg:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/AIZOOTech/mask-detection-web-demo/b4425bc5fa5a75bd52e17a53a1a0b0243f8e5e7a/images/demo.jpg


--------------------------------------------------------------------------------
/images/result.png:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/AIZOOTech/mask-detection-web-demo/b4425bc5fa5a75bd52e17a53a1a0b0243f8e5e7a/images/result.png


--------------------------------------------------------------------------------
/index.html:
--------------------------------------------------------------------------------
 1 |  <!--
 2 | Copyright (c) 2018 ml5
 3 | 
 4 | This software is released under the MIT License.
 5 | https://opensource.org/licenses/MIT
 6 | -->
 7 | 
 8 | <html>
 9 | <head>
10 |   <meta charset="UTF-8">
11 |   <link rel="stylesheet" type="text/css" href="css/style.css">
12 |   <script src="js/tfjs.min.js"></script>
13 | </head>
14 | 
15 | <body onresize="windowResized()">
16 |   <a href="https://aizoo.com"><img src=images/aizoocom.png height="30px"></a>
17 |   <h1 class='title'>借助TensorFlow.js</h1>
18 |   <h1 class='title'>让人脸口罩识别跑在您本地浏览器里面</h1>
19 |   <h5 class='title' id='tips'> 注意：部分手机拍摄的图片，读取出来是旋转90°，需要旋转正，为避免本demo过于复杂，本demo不对这种case进行处理</h3>
20 |   <div id="container">
21 |     <img src="images/demo.jpg" id="image" class="center">
22 |     <div id="text-container">
23 |       <p id="instruction" class="green-color">
24 |       <button id="uploader-btn" onClick="clickUploader()">打开本地图片</button>
25 |       <p class="green-color "> 或者将图片拖到这里 📂,  然后下方会显示结果</p>
26 |     
27 |       <p id="warning" class="result-color"></p>
28 |       <p class="result-color">
29 |         检测结果如下：
30 |       </p>
31 |       <input name="imgFile" type="file" id="fileUploader" accept="image/*" onChange="handleFiles()">
32 |     </div>
33 |     <canvas id="canvas"></canvas>
34 |   </div>
35 |   <script src="js/detection.js"></script>
36 |   <script src="js/index.js"></script>
37 | </body>
38 | </html>
39 | 


--------------------------------------------------------------------------------
/js/detection.js:
--------------------------------------------------------------------------------
  1 | // decode the output according to anchors
  2 | function decodeBBox(anchors, rawOutput, variances=[0.1,0.1,0.2,0.2]) {
  3 |     const [anchorXmin, anchorYmin, anchorXmax, anchorYmax] = tf.split(anchors, [1,1,1,1], -1);
  4 |     const anchorCX = tf.div(tf.add(anchorXmin, anchorXmax), 2);
  5 |     const anchorCY = tf.div(tf.add(anchorYmin, anchorYmax), 2);
  6 | 
  7 |     const anchorW = tf.sub(anchorXmax, anchorXmin);
  8 |     const anchorH = tf.sub(anchorYmax, anchorYmin);
  9 | 
 10 |     const rawOutputScale = tf.mul(rawOutput, tf.tensor(variances));
 11 |     const [rawOutputCX, rawOutputCY, rawOutputW, rawOutputH] = tf.split(rawOutputScale, [1,1,1,1], -1);
 12 |     const predictCX = tf.add(tf.mul(rawOutputCX, anchorW), anchorCX);
 13 |     const predictCY = tf.add(tf.mul(rawOutputCY,anchorH), anchorCY);
 14 |     const predictW = tf.mul(tf.exp(rawOutputW), anchorW);
 15 |     const predictH = tf.mul(tf.exp(rawOutputH), anchorH);
 16 |     const predictXmin = tf.sub(predictCX, tf.div(predictW, 2));
 17 |     const predictYmin = tf.sub(predictCY, tf.div(predictH, 2));
 18 |     const predictXmax = tf.add(predictCX, tf.div(predictW, 2));
 19 |     const predictYmax = tf.add(predictCY, tf.div(predictH, 2));
 20 |     // eslint-disable-next-line
 21 |     const predictBBox = tf.concat([predictYmin, predictXmin, predictYmax, predictXmax],-1);
 22 |     return predictBBox
 23 | }
 24 | 
 25 | // generate anchors
 26 | function anchorGenerator(featureMapSizes, anchorSizes, anchorRatios) {
 27 |     let anchorBBoxes = [];
 28 |     // eslint-disable-next-line
 29 |     featureMapSizes.map((featureSize, idx) =>{
 30 |         const cx = tf.div(tf.add(tf.linspace(0, featureSize[0] - 1, featureSize[0]) , 0.5) , featureSize[0]);
 31 |         const cy = tf.div(tf.add(tf.linspace(0, featureSize[1] - 1, featureSize[1]) , 0.5) , featureSize[1]);
 32 |         const cxGrid = tf.matMul(tf.ones([featureSize[1], 1]), cx.reshape([1,featureSize[0]]));
 33 |         const cyGrid = tf.matMul(cy.reshape([featureSize[1], 1]), tf.ones([1, featureSize[0]]));
 34 |         // eslint-disable-next-line
 35 |         const cxGridExpend = tf.expandDims(cxGrid, -1);
 36 |         // eslint-disable-next-line
 37 |         const cyGridExpend = tf.expandDims(cyGrid, -1);
 38 |         // eslint-disable-next-line
 39 |         const center = tf.concat([cxGridExpend, cyGridExpend], -1);
 40 |         const numAnchors = anchorSizes[idx].length + anchorRatios[idx].length -1;
 41 |         const centerTiled = tf.tile(center, [1, 1, 2*numAnchors]);
 42 |         // eslint-disable-next-line
 43 |         let anchorWidthHeights = [];
 44 |         
 45 |         // eslint-disable-next-line
 46 |         for (const scale of anchorSizes[idx]) {
 47 |             const ratio = anchorRatios[idx][0];
 48 |             const width = scale * Math.sqrt(ratio);
 49 |             const height = scale / Math.sqrt(ratio);
 50 | 
 51 |             const halfWidth = width /  2;
 52 |             const halfHeight = height / 2;
 53 |             anchorWidthHeights.push(-halfWidth, -halfHeight, halfWidth, halfHeight);
 54 |             // width = tf.mul(scale, tf.sqrt(ratio));
 55 |             // height = tf.div(scale, tf.sqrt(ratio));
 56 | 
 57 |             // halfWidth = tf.div(width, 2);
 58 |             // halfHeight = tf.div(height, 2);
 59 |             // anchorWidthHeights.push(tf.neg(halfWidth), tf.neg(halfWidth), halfWidth, halfHeight);
 60 |         }
 61 | 
 62 |         // eslint-disable-next-line
 63 |         for ( const ratio of anchorRatios[idx].slice(1)) {
 64 |             const scale = anchorSizes[idx][0];
 65 |             const width = scale * Math.sqrt(ratio);
 66 |             const height = scale / Math.sqrt(ratio);
 67 |             const halfWidth = width /  2;
 68 |             const halfHeight = height / 2;
 69 |             anchorWidthHeights.push(-halfWidth, -halfHeight, halfWidth, halfHeight);
 70 |         }
 71 |         const bboxCoord = tf.add(centerTiled , tf.tensor(anchorWidthHeights));
 72 |         const bboxCoordReshape = bboxCoord.reshape([-1, 4]);
 73 |         anchorBBoxes.push(bboxCoordReshape);
 74 |     })
 75 |     // eslint-disable-next-line
 76 |     anchorBBoxes = tf.concat(anchorBBoxes, 0);
 77 |     return anchorBBoxes;
 78 | }
 79 | 
 80 | //  nms function
 81 | function nonMaxSuppression(bboxes, confidences, confThresh, iouThresh, width, height, maxOutputSize=100) {
 82 |     const bboxMaxFlag = tf.argMax(confidences, -1);
 83 |     const bboxConf = tf.max(confidences, -1);
 84 |     const keepIndices = tf.image.nonMaxSuppression(bboxes, bboxConf, maxOutputSize, iouThresh, confThresh);
 85 |     // eslint-disable-next-line
 86 |     let results = []
 87 |     const keepIndicesData = keepIndices.dataSync();
 88 |     const bboxConfData = bboxConf.dataSync();
 89 |     const bboxMaxFlagData = bboxMaxFlag.dataSync();
 90 |     const bboxesData = bboxes.dataSync();
 91 |     // eslint-disable-next-line
 92 |     keepIndicesData.map((idx) => {
 93 |         const xmin = Math.round(Math.max(bboxesData[4*idx + 1] * width, 0));
 94 |         const ymin = Math.round(Math.max(bboxesData[4*idx + 0] * height, 0));
 95 |         const xmax = Math.round(Math.min(bboxesData[4*idx+3] * width, width))
 96 |         const ymax = Math.round(Math.min(bboxesData[4*idx + 2] * height, height));
 97 |         results.push([[xmin, ymin, xmax, ymax],
 98 |             bboxMaxFlagData[idx], bboxConfData[idx]])
 99 |     });
100 |     return results;
101 | }
102 | 
103 | 
104 | async function loadModel() {
105 |     model = await tf.loadLayersModel('./tfjs-models/model.json');
106 |     return model;
107 | }
108 | 
109 | async function detect(imgToPredict) {
110 |     const detectionResults = tf.tidy(() => {
111 |         // eslint-disable-next-line
112 |         const width = imgToPredict.width;
113 |         // eslint-disable-next-line
114 |         const height = imgToPredict.height;
115 |         let img = tf.browser.fromPixels(imgToPredict);
116 |         img = tf.image.resizeBilinear(img, [260, 260]);
117 |         img = img.expandDims(0).toFloat().div(tf.scalar(255));
118 |         const [rawBBoxes, rawConfidences] = model.predict(img);
119 |         const bboxes = decodeBBox(anchors, tf.squeeze(rawBBoxes));
120 |         const Results = nonMaxSuppression(bboxes, tf.squeeze(rawConfidences), 0.5, 0.5,  width, height );
121 |         return Results;
122 |     })
123 |     return detectionResults;
124 | }
125 | 
126 | featureMapSizes = [[33, 33], [17, 17], [9, 9], [5, 5], [3,3]];
127 | anchorSizes =  [[0.04, 0.056], [0.08, 0.11], [0.16, 0.22], [0.32, 0.45], [0.64, 0.72]];
128 | anchorRatios = [[1, 0.62, 0.42], [1, 0.62, 0.42], [1, 0.62, 0.42], [1, 0.62, 0.42], [1, 0.62, 0.42]];
129 | 
130 | let anchors = anchorGenerator(featureMapSizes, anchorSizes, anchorRatios);
131 | 


--------------------------------------------------------------------------------
/js/index.js:
--------------------------------------------------------------------------------
  1 | 
  2 | const image = document.getElementById('image'); 
  3 | const canvas = document.getElementById('canvas');
  4 | const dropContainer = document.getElementById('container');
  5 | const warning = document.getElementById('warning');
  6 | const fileInput = document.getElementById('fileUploader');
  7 | 
  8 | const id2class = {0:"有口罩", 1:"无口罩"};
  9 | let model;
 10 | 
 11 | function preventDefaults(e) {
 12 |   e.preventDefault()
 13 |   e.stopPropagation()
 14 | };
 15 | 
 16 | function windowResized() {
 17 |   let windowW = window.innerWidth;
 18 |   if (windowW < 480 && windowW >= 200) {
 19 |     dropContainer.style.display = 'block';
 20 |   } else if (windowW < 200) {
 21 |     dropContainer.style.display = 'none';
 22 |   } else {
 23 |     dropContainer.style.display = 'block';
 24 |   }
 25 | }
 26 | 
 27 | ['dragenter', 'dragover'].forEach(eventName => {
 28 |   dropContainer.addEventListener(eventName, e => dropContainer.classList.add('highlight'), false)
 29 | });
 30 | 
 31 | ['dragleave', 'drop'].forEach(eventName => {
 32 |   dropContainer.addEventListener(eventName, e => dropContainer.classList.remove('highlight'), false)
 33 | });
 34 | 
 35 | ['dragenter', 'dragover', 'dragleave', 'drop'].forEach(eventName => {
 36 |   dropContainer.addEventListener(eventName, preventDefaults, false)
 37 | });
 38 | 
 39 | dropContainer.addEventListener('drop', gotImage, false)
 40 | 
 41 | 
 42 | function gotImage(e) {
 43 |   const dt = e.dataTransfer;
 44 |   const files = dt.files;
 45 |   if (files.length > 1) {
 46 |     console.error('upload only one file');
 47 |   }
 48 |   const file = files[0];
 49 |   const imageType = /image.*/;
 50 |   if (file.type.match(imageType)) {
 51 |     warning.innerHTML = '';
 52 |     const reader = new FileReader();
 53 |     reader.readAsDataURL(file);
 54 |     reader.onloadend = () => {
 55 |       image.src = reader.result;
 56 |       setTimeout(detectImage, 100);
 57 |     }
 58 |   } else {
 59 |     image.src = 'images/demo.jpg';
 60 |     setTimeout(detectImage, 100);
 61 |     warning.innerHTML = 'Please drop an image file.'
 62 |   }
 63 | }
 64 | 
 65 | function handleFiles() {
 66 |   const curFiles = fileInput.files;
 67 |   if (curFiles.length === 0) {
 68 |     image.src = 'images/demo.jpg';
 69 |     setTimeout(detectImage, 100);
 70 |     warning.innerHTML = 'No image selected for upload';
 71 |   } else {
 72 |     image.src = window.URL.createObjectURL(curFiles[0]);
 73 |     warning.innerHTML = '';
 74 |     setTimeout(detectImage, 100);
 75 |   }
 76 | }
 77 | 
 78 | function clickUploader() {
 79 |   fileInput.click();
 80 | }
 81 | 
 82 | // 检测人脸和口罩
 83 | function detectImage() {
 84 |   detect(image).then((results) => {
 85 |     canvas.width = image.width;
 86 |     canvas.height = image.height;
 87 |     ctx = canvas.getContext('2d');
 88 |     ctx.clearRect(0, 0, canvas.width, canvas.height);
 89 |     ctx.drawImage(image, 0, 0);
 90 |     for(bboxInfo of results) {
 91 |       bbox = bboxInfo[0];
 92 |       classID = bboxInfo[1];
 93 |       score = bboxInfo[2];
 94 | 
 95 |       ctx.beginPath();
 96 |       ctx.lineWidth="4";
 97 |       if (classID == 0) {
 98 |           ctx.strokeStyle="green";
 99 |           ctx.fillStyle="green";
100 |       } else {
101 |           ctx.strokeStyle="red";
102 |           ctx.fillStyle="red";
103 |       }
104 |       
105 |       ctx.rect(bbox[0], bbox[1], bbox[2] - bbox[0], bbox[3] - bbox[1]);
106 |       ctx.stroke();
107 |       
108 |       ctx.font="30px Arial";
109 |       
110 |       let content = id2class[classID] + " " + score.toFixed(2);
111 |       ctx.fillText(content, bbox[0], bbox[1] < 20 ? bbox[1] + 30 : bbox[1]-5);
112 |   }
113 |   })
114 | }
115 | 
116 | // 初始化函数
117 | async function setup() {
118 |   await loadModel();
119 |   // Make a detection with the default image
120 |   detectImage();
121 | }
122 | 
123 | setup();
124 | 


--------------------------------------------------------------------------------
/tfjs-models/group1-shard1of1.bin:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/AIZOOTech/mask-detection-web-demo/b4425bc5fa5a75bd52e17a53a1a0b0243f8e5e7a/tfjs-models/group1-shard1of1.bin


--------------------------------------------------------------------------------
/tfjs-models/model.json:
--------------------------------------------------------------------------------
1 | {"format": "layers-model", "generatedBy": "keras v2.2.4", "convertedBy": "TensorFlow.js Converter v1.5.2", "modelTopology": {"keras_version": "2.2.4", "backend": "tensorflow", "model_config": {"class_name": "Model", "config": {"layers": [{"class_name": "InputLayer", "config": {"dtype": "float32", "batch_input_shape": [null, 260, 260, 3], "name": "data", "sparse": false}, "inbound_nodes": [], "name": "data"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_0", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 32, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["data", 0, 0, {}]]], "name": "conv2d_0"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_0_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_0", 0, 0, {}]]], "name": "conv2d_0_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_0_activation"}, "inbound_nodes": [[["conv2d_0_bn", 0, 0, {}]]], "name": "conv2d_0_activation"}, {"class_name": "MaxPooling2D", "config": {"name": "maxpool2d_0", "trainable": true, "data_format": "channels_last", "pool_size": [2, 2], "padding": "same", "strides": [2, 2]}, "inbound_nodes": [[["conv2d_0_activation", 0, 0, {}]]], "name": "maxpool2d_0"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_1", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["maxpool2d_0", 0, 0, {}]]], "name": "conv2d_1"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_1_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_1", 0, 0, {}]]], "name": "conv2d_1_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_1_activation"}, "inbound_nodes": [[["conv2d_1_bn", 0, 0, {}]]], "name": "conv2d_1_activation"}, {"class_name": "MaxPooling2D", "config": {"name": "maxpool2d_1", "trainable": true, "data_format": "channels_last", "pool_size": [2, 2], "padding": "same", "strides": [2, 2]}, "inbound_nodes": [[["conv2d_1_activation", 0, 0, {}]]], "name": "maxpool2d_1"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_2", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["maxpool2d_1", 0, 0, {}]]], "name": "conv2d_2"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_2_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_2", 0, 0, {}]]], "name": "conv2d_2_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_2_activation"}, "inbound_nodes": [[["conv2d_2_bn", 0, 0, {}]]], "name": "conv2d_2_activation"}, {"class_name": "MaxPooling2D", "config": {"name": "maxpool2d_2", "trainable": true, "data_format": "channels_last", "pool_size": [2, 2], "padding": "same", "strides": [2, 2]}, "inbound_nodes": [[["conv2d_2_activation", 0, 0, {}]]], "name": "maxpool2d_2"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_3", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["maxpool2d_2", 0, 0, {}]]], "name": "conv2d_3"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_3_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_3", 0, 0, {}]]], "name": "conv2d_3_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_3_activation"}, "inbound_nodes": [[["conv2d_3_bn", 0, 0, {}]]], "name": "conv2d_3_activation"}, {"class_name": "MaxPooling2D", "config": {"name": "maxpool2d_3", "trainable": true, "data_format": "channels_last", "pool_size": [2, 2], "padding": "same", "strides": [2, 2]}, "inbound_nodes": [[["conv2d_3_activation", 0, 0, {}]]], "name": "maxpool2d_3"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_4", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 128, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["maxpool2d_3", 0, 0, {}]]], "name": "conv2d_4"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_4_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_4", 0, 0, {}]]], "name": "conv2d_4_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_4_activation"}, "inbound_nodes": [[["conv2d_4_bn", 0, 0, {}]]], "name": "conv2d_4_activation"}, {"class_name": "MaxPooling2D", "config": {"name": "maxpool2d_4", "trainable": true, "data_format": "channels_last", "pool_size": [2, 2], "padding": "same", "strides": [2, 2]}, "inbound_nodes": [[["conv2d_4_activation", 0, 0, {}]]], "name": "maxpool2d_4"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_5", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 128, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["maxpool2d_4", 0, 0, {}]]], "name": "conv2d_5"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_5_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_5", 0, 0, {}]]], "name": "conv2d_5_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_5_activation"}, "inbound_nodes": [[["conv2d_5_bn", 0, 0, {}]]], "name": "conv2d_5_activation"}, {"class_name": "MaxPooling2D", "config": {"name": "maxpool2d_5", "trainable": true, "data_format": "channels_last", "pool_size": [2, 2], "padding": "same", "strides": [2, 2]}, "inbound_nodes": [[["conv2d_5_activation", 0, 0, {}]]], "name": "maxpool2d_5"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_6", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["maxpool2d_5", 0, 0, {}]]], "name": "conv2d_6"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_6_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_6", 0, 0, {}]]], "name": "conv2d_6_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_6_activation"}, "inbound_nodes": [[["conv2d_6_bn", 0, 0, {}]]], "name": "conv2d_6_activation"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "conv2d_7", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "valid", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_6_activation", 0, 0, {}]]], "name": "conv2d_7"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "conv2d_7_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["conv2d_7", 0, 0, {}]]], "name": "conv2d_7_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "conv2d_7_activation"}, "inbound_nodes": [[["conv2d_7_bn", 0, 0, {}]]], "name": "conv2d_7_activation"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "cls_0_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_3_activation", 0, 0, {}]]], "name": "cls_0_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "cls_1_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_4_activation", 0, 0, {}]]], "name": "cls_1_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "cls_2_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_5_activation", 0, 0, {}]]], "name": "cls_2_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "cls_3_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_6_activation", 0, 0, {}]]], "name": "cls_3_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "cls_4_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_7_activation", 0, 0, {}]]], "name": "cls_4_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "loc_0_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_3_activation", 0, 0, {}]]], "name": "loc_0_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "loc_1_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_4_activation", 0, 0, {}]]], "name": "loc_1_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "loc_2_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_5_activation", 0, 0, {}]]], "name": "loc_2_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "loc_3_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_6_activation", 0, 0, {}]]], "name": "loc_3_insert_conv2d"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "normal", "scale": 2.0, "seed": null, "mode": "fan_in"}}, "name": "loc_4_insert_conv2d", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": {"class_name": "L1L2", "config": {"l2": 9.999999747378752e-05, "l1": 0.0}}, "filters": 64, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": false, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["conv2d_7_activation", 0, 0, {}]]], "name": "loc_4_insert_conv2d"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "cls_0_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["cls_0_insert_conv2d", 0, 0, {}]]], "name": "cls_0_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "cls_1_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["cls_1_insert_conv2d", 0, 0, {}]]], "name": "cls_1_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "cls_2_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["cls_2_insert_conv2d", 0, 0, {}]]], "name": "cls_2_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "cls_3_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["cls_3_insert_conv2d", 0, 0, {}]]], "name": "cls_3_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "cls_4_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["cls_4_insert_conv2d", 0, 0, {}]]], "name": "cls_4_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "loc_0_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["loc_0_insert_conv2d", 0, 0, {}]]], "name": "loc_0_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "loc_1_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["loc_1_insert_conv2d", 0, 0, {}]]], "name": "loc_1_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "loc_2_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["loc_2_insert_conv2d", 0, 0, {}]]], "name": "loc_2_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "loc_3_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["loc_3_insert_conv2d", 0, 0, {}]]], "name": "loc_3_insert_conv2d_bn"}, {"class_name": "BatchNormalization", "config": {"beta_constraint": null, "gamma_initializer": {"class_name": "Ones", "config": {}}, "moving_mean_initializer": {"class_name": "Zeros", "config": {}}, "name": "loc_4_insert_conv2d_bn", "epsilon": 0.001, "trainable": true, "moving_variance_initializer": {"class_name": "Ones", "config": {}}, "beta_initializer": {"class_name": "Zeros", "config": {}}, "scale": true, "axis": 3, "gamma_constraint": null, "gamma_regularizer": null, "beta_regularizer": null, "momentum": 0.99, "center": true}, "inbound_nodes": [[["loc_4_insert_conv2d", 0, 0, {}]]], "name": "loc_4_insert_conv2d_bn"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "cls_0_insert_conv2d_activation"}, "inbound_nodes": [[["cls_0_insert_conv2d_bn", 0, 0, {}]]], "name": "cls_0_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "cls_1_insert_conv2d_activation"}, "inbound_nodes": [[["cls_1_insert_conv2d_bn", 0, 0, {}]]], "name": "cls_1_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "cls_2_insert_conv2d_activation"}, "inbound_nodes": [[["cls_2_insert_conv2d_bn", 0, 0, {}]]], "name": "cls_2_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "cls_3_insert_conv2d_activation"}, "inbound_nodes": [[["cls_3_insert_conv2d_bn", 0, 0, {}]]], "name": "cls_3_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "cls_4_insert_conv2d_activation"}, "inbound_nodes": [[["cls_4_insert_conv2d_bn", 0, 0, {}]]], "name": "cls_4_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "loc_0_insert_conv2d_activation"}, "inbound_nodes": [[["loc_0_insert_conv2d_bn", 0, 0, {}]]], "name": "loc_0_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "loc_1_insert_conv2d_activation"}, "inbound_nodes": [[["loc_1_insert_conv2d_bn", 0, 0, {}]]], "name": "loc_1_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "loc_2_insert_conv2d_activation"}, "inbound_nodes": [[["loc_2_insert_conv2d_bn", 0, 0, {}]]], "name": "loc_2_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "loc_3_insert_conv2d_activation"}, "inbound_nodes": [[["loc_3_insert_conv2d_bn", 0, 0, {}]]], "name": "loc_3_insert_conv2d_activation"}, {"class_name": "Activation", "config": {"activation": "relu", "trainable": true, "name": "loc_4_insert_conv2d_activation"}, "inbound_nodes": [[["loc_4_insert_conv2d_bn", 0, 0, {}]]], "name": "loc_4_insert_conv2d_activation"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "cls_0_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 8, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["cls_0_insert_conv2d_activation", 0, 0, {}]]], "name": "cls_0_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "cls_1_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 8, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["cls_1_insert_conv2d_activation", 0, 0, {}]]], "name": "cls_1_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "cls_2_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 8, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["cls_2_insert_conv2d_activation", 0, 0, {}]]], "name": "cls_2_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "cls_3_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 8, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["cls_3_insert_conv2d_activation", 0, 0, {}]]], "name": "cls_3_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "cls_4_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 8, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["cls_4_insert_conv2d_activation", 0, 0, {}]]], "name": "cls_4_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "loc_0_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 16, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["loc_0_insert_conv2d_activation", 0, 0, {}]]], "name": "loc_0_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "loc_1_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 16, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["loc_1_insert_conv2d_activation", 0, 0, {}]]], "name": "loc_1_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "loc_2_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 16, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["loc_2_insert_conv2d_activation", 0, 0, {}]]], "name": "loc_2_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "loc_3_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 16, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["loc_3_insert_conv2d_activation", 0, 0, {}]]], "name": "loc_3_conv"}, {"class_name": "Conv2D", "config": {"kernel_constraint": null, "kernel_initializer": {"class_name": "VarianceScaling", "config": {"distribution": "uniform", "scale": 1.0, "seed": null, "mode": "fan_avg"}}, "name": "loc_4_conv", "bias_regularizer": null, "bias_constraint": null, "activation": "linear", "trainable": true, "data_format": "channels_last", "padding": "same", "strides": [1, 1], "dilation_rate": [1, 1], "kernel_regularizer": null, "filters": 16, "bias_initializer": {"class_name": "Zeros", "config": {}}, "use_bias": true, "activity_regularizer": null, "kernel_size": [3, 3]}, "inbound_nodes": [[["loc_4_insert_conv2d_activation", 0, 0, {}]]], "name": "loc_4_conv"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 2], "trainable": true, "name": "cls_0_reshape"}, "inbound_nodes": [[["cls_0_conv", 0, 0, {}]]], "name": "cls_0_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 2], "trainable": true, "name": "cls_1_reshape"}, "inbound_nodes": [[["cls_1_conv", 0, 0, {}]]], "name": "cls_1_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 2], "trainable": true, "name": "cls_2_reshape"}, "inbound_nodes": [[["cls_2_conv", 0, 0, {}]]], "name": "cls_2_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 2], "trainable": true, "name": "cls_3_reshape"}, "inbound_nodes": [[["cls_3_conv", 0, 0, {}]]], "name": "cls_3_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 2], "trainable": true, "name": "cls_4_reshape"}, "inbound_nodes": [[["cls_4_conv", 0, 0, {}]]], "name": "cls_4_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 4], "trainable": true, "name": "loc_0_reshape"}, "inbound_nodes": [[["loc_0_conv", 0, 0, {}]]], "name": "loc_0_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 4], "trainable": true, "name": "loc_1_reshape"}, "inbound_nodes": [[["loc_1_conv", 0, 0, {}]]], "name": "loc_1_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 4], "trainable": true, "name": "loc_2_reshape"}, "inbound_nodes": [[["loc_2_conv", 0, 0, {}]]], "name": "loc_2_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 4], "trainable": true, "name": "loc_3_reshape"}, "inbound_nodes": [[["loc_3_conv", 0, 0, {}]]], "name": "loc_3_reshape"}, {"class_name": "Reshape", "config": {"target_shape": [-1, 4], "trainable": true, "name": "loc_4_reshape"}, "inbound_nodes": [[["loc_4_conv", 0, 0, {}]]], "name": "loc_4_reshape"}, {"class_name": "Activation", "config": {"activation": "sigmoid", "trainable": true, "name": "cls_0_activation"}, "inbound_nodes": [[["cls_0_reshape", 0, 0, {}]]], "name": "cls_0_activation"}, {"class_name": "Activation", "config": {"activation": "sigmoid", "trainable": true, "name": "cls_1_activation"}, "inbound_nodes": [[["cls_1_reshape", 0, 0, {}]]], "name": "cls_1_activation"}, {"class_name": "Activation", "config": {"activation": "sigmoid", "trainable": true, "name": "cls_2_activation"}, "inbound_nodes": [[["cls_2_reshape", 0, 0, {}]]], "name": "cls_2_activation"}, {"class_name": "Activation", "config": {"activation": "sigmoid", "trainable": true, "name": "cls_3_activation"}, "inbound_nodes": [[["cls_3_reshape", 0, 0, {}]]], "name": "cls_3_activation"}, {"class_name": "Activation", "config": {"activation": "sigmoid", "trainable": true, "name": "cls_4_activation"}, "inbound_nodes": [[["cls_4_reshape", 0, 0, {}]]], "name": "cls_4_activation"}, {"class_name": "Concatenate", "config": {"trainable": true, "name": "loc_branch_concat", "axis": 1}, "inbound_nodes": [[["loc_0_reshape", 0, 0, {}], ["loc_1_reshape", 0, 0, {}], ["loc_2_reshape", 0, 0, {}], ["loc_3_reshape", 0, 0, {}], ["loc_4_reshape", 0, 0, {}]]], "name": "loc_branch_concat"}, {"class_name": "Concatenate", "config": {"trainable": true, "name": "cls_branch_concat", "axis": 1}, "inbound_nodes": [[["cls_0_activation", 0, 0, {}], ["cls_1_activation", 0, 0, {}], ["cls_2_activation", 0, 0, {}], ["cls_3_activation", 0, 0, {}], ["cls_4_activation", 0, 0, {}]]], "name": "cls_branch_concat"}], "input_layers": [["data", 0, 0]], "output_layers": [["loc_branch_concat", 0, 0], ["cls_branch_concat", 0, 0]], "name": "model_2"}}, "training_config": {"metrics": [], "loss": ["loc_loss", "cls_loss"], "optimizer_config": {"class_name": "Adam", "config": {"beta_1": 0.8999999761581421, "beta_2": 0.9990000128746033, "decay": 0.0, "epsilon": 1e-07, "amsgrad": false, "lr": 0.0004923854139633477}}, "loss_weights": [1, 1], "sample_weight_mode": null}}, "weightsManifest": [{"paths": ["group1-shard1of1.bin"], "weights": [{"name": "cls_0_conv/kernel", "shape": [3, 3, 64, 8], "dtype": "float32"}, {"name": "cls_0_conv/bias", "shape": [8], "dtype": "float32"}, {"name": "cls_0_insert_conv2d/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "cls_0_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "cls_0_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "cls_0_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "cls_0_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "cls_1_conv/kernel", "shape": [3, 3, 64, 8], "dtype": "float32"}, {"name": "cls_1_conv/bias", "shape": [8], "dtype": "float32"}, {"name": "cls_1_insert_conv2d/kernel", "shape": [3, 3, 128, 64], "dtype": "float32"}, {"name": "cls_1_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "cls_1_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "cls_1_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "cls_1_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "cls_2_conv/kernel", "shape": [3, 3, 64, 8], "dtype": "float32"}, {"name": "cls_2_conv/bias", "shape": [8], "dtype": "float32"}, {"name": "cls_2_insert_conv2d/kernel", "shape": [3, 3, 128, 64], "dtype": "float32"}, {"name": "cls_2_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "cls_2_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "cls_2_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "cls_2_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "cls_3_conv/kernel", "shape": [3, 3, 64, 8], "dtype": "float32"}, {"name": "cls_3_conv/bias", "shape": [8], "dtype": "float32"}, {"name": "cls_3_insert_conv2d/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "cls_3_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "cls_3_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "cls_3_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "cls_3_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "cls_4_conv/kernel", "shape": [3, 3, 64, 8], "dtype": "float32"}, {"name": "cls_4_conv/bias", "shape": [8], "dtype": "float32"}, {"name": "cls_4_insert_conv2d/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "cls_4_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "cls_4_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "cls_4_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "cls_4_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "conv2d_0/kernel", "shape": [3, 3, 3, 32], "dtype": "float32"}, {"name": "conv2d_0_bn/gamma", "shape": [32], "dtype": "float32"}, {"name": "conv2d_0_bn/beta", "shape": [32], "dtype": "float32"}, {"name": "conv2d_0_bn/moving_mean", "shape": [32], "dtype": "float32"}, {"name": "conv2d_0_bn/moving_variance", "shape": [32], "dtype": "float32"}, {"name": "conv2d_1/kernel", "shape": [3, 3, 32, 64], "dtype": "float32"}, {"name": "conv2d_1_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "conv2d_1_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "conv2d_1_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "conv2d_1_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "conv2d_2/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "conv2d_2_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "conv2d_2_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "conv2d_2_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "conv2d_2_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "conv2d_3/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "conv2d_3_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "conv2d_3_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "conv2d_3_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "conv2d_3_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "conv2d_4/kernel", "shape": [3, 3, 64, 128], "dtype": "float32"}, {"name": "conv2d_4_bn/gamma", "shape": [128], "dtype": "float32"}, {"name": "conv2d_4_bn/beta", "shape": [128], "dtype": "float32"}, {"name": "conv2d_4_bn/moving_mean", "shape": [128], "dtype": "float32"}, {"name": "conv2d_4_bn/moving_variance", "shape": [128], "dtype": "float32"}, {"name": "conv2d_5/kernel", "shape": [3, 3, 128, 128], "dtype": "float32"}, {"name": "conv2d_5_bn/gamma", "shape": [128], "dtype": "float32"}, {"name": "conv2d_5_bn/beta", "shape": [128], "dtype": "float32"}, {"name": "conv2d_5_bn/moving_mean", "shape": [128], "dtype": "float32"}, {"name": "conv2d_5_bn/moving_variance", "shape": [128], "dtype": "float32"}, {"name": "conv2d_6/kernel", "shape": [3, 3, 128, 64], "dtype": "float32"}, {"name": "conv2d_6_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "conv2d_6_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "conv2d_6_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "conv2d_6_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "conv2d_7/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "conv2d_7_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "conv2d_7_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "conv2d_7_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "conv2d_7_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "loc_0_conv/kernel", "shape": [3, 3, 64, 16], "dtype": "float32"}, {"name": "loc_0_conv/bias", "shape": [16], "dtype": "float32"}, {"name": "loc_0_insert_conv2d/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "loc_0_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "loc_0_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "loc_0_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "loc_0_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "loc_1_conv/kernel", "shape": [3, 3, 64, 16], "dtype": "float32"}, {"name": "loc_1_conv/bias", "shape": [16], "dtype": "float32"}, {"name": "loc_1_insert_conv2d/kernel", "shape": [3, 3, 128, 64], "dtype": "float32"}, {"name": "loc_1_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "loc_1_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "loc_1_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "loc_1_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "loc_2_conv/kernel", "shape": [3, 3, 64, 16], "dtype": "float32"}, {"name": "loc_2_conv/bias", "shape": [16], "dtype": "float32"}, {"name": "loc_2_insert_conv2d/kernel", "shape": [3, 3, 128, 64], "dtype": "float32"}, {"name": "loc_2_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "loc_2_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "loc_2_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "loc_2_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "loc_3_conv/kernel", "shape": [3, 3, 64, 16], "dtype": "float32"}, {"name": "loc_3_conv/bias", "shape": [16], "dtype": "float32"}, {"name": "loc_3_insert_conv2d/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "loc_3_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "loc_3_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "loc_3_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "loc_3_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}, {"name": "loc_4_conv/kernel", "shape": [3, 3, 64, 16], "dtype": "float32"}, {"name": "loc_4_conv/bias", "shape": [16], "dtype": "float32"}, {"name": "loc_4_insert_conv2d/kernel", "shape": [3, 3, 64, 64], "dtype": "float32"}, {"name": "loc_4_insert_conv2d_bn/gamma", "shape": [64], "dtype": "float32"}, {"name": "loc_4_insert_conv2d_bn/beta", "shape": [64], "dtype": "float32"}, {"name": "loc_4_insert_conv2d_bn/moving_mean", "shape": [64], "dtype": "float32"}, {"name": "loc_4_insert_conv2d_bn/moving_variance", "shape": [64], "dtype": "float32"}]}]}


--------------------------------------------------------------------------------