                        課程
                    
                        /前端開發(fā)
                        
                            /JavaScript
                        
                        /Tensorflow.js 實(shí)現(xiàn)垃圾分類

實(shí)際通過tf.loadLayersModel拿到的mobilenet的模型和課程中不一樣要怎么處理呢？后續(xù)出現(xiàn)報(bào)錯(cuò)

http://ai-sample.oss-cn-hangzhou.aliyuncs.com/pipcook/models/mobilenet/web_model/model.json

下面是調(diào)用

const?mobilenet?=?await?tf.loadLayersModel(MOBILENET_URL)

????mobilenet.summary()

打印的數(shù)據(jù)如下

_________________________________________________________________

Layer (type)? ? ? ? ? ? ? ? ?Output shape? ? ? ? ? ? ? Param #? ?

=================================================================

input_1 (InputLayer)? ? ? ? ?[null,null,null,3]? ? ? ? 0

_________________________________________________________________

conv1 (Conv2D)? ? ? ? ? ? ? ?[null,null,null,32]? ? ? ?864

_________________________________________________________________

conv1_bn (BatchNormalization [null,null,null,32]? ? ? ?128

_________________________________________________________________

conv1_relu (ReLU)? ? ? ? ? ? [null,null,null,32]? ? ? ?0

_________________________________________________________________

conv_dw_1 (DepthwiseConv2D)? [null,null,null,32]? ? ? ?288

_________________________________________________________________

conv_dw_1_bn (BatchNormaliza [null,null,null,32]? ? ? ?128

_________________________________________________________________

conv_dw_1_relu (ReLU)? ? ? ? [null,null,null,32]? ? ? ?0

_________________________________________________________________

conv_pw_1 (Conv2D)? ? ? ? ? ?[null,null,null,64]? ? ? ?2048

_________________________________________________________________

conv_pw_1_bn (BatchNormaliza [null,null,null,64]? ? ? ?256

_________________________________________________________________

conv_pw_1_relu (ReLU)? ? ? ? [null,null,null,64]? ? ? ?0

_________________________________________________________________

conv_pad_2 (ZeroPadding2D)? ?[null,null,null,64]? ? ? ?0

_________________________________________________________________

conv_dw_2 (DepthwiseConv2D)? [null,null,null,64]? ? ? ?576

_________________________________________________________________

conv_dw_2_bn (BatchNormaliza [null,null,null,64]? ? ? ?256

_________________________________________________________________

conv_dw_2_relu (ReLU)? ? ? ? [null,null,null,64]? ? ? ?0

_________________________________________________________________

conv_pw_2 (Conv2D)? ? ? ? ? ?[null,null,null,128]? ? ? 8192

_________________________________________________________________

conv_pw_2_bn (BatchNormaliza [null,null,null,128]? ? ? 512

_________________________________________________________________

conv_pw_2_relu (ReLU)? ? ? ? [null,null,null,128]? ? ? 0

_________________________________________________________________

conv_dw_3 (DepthwiseConv2D)? [null,null,null,128]? ? ? 1152

_________________________________________________________________

conv_dw_3_bn (BatchNormaliza [null,null,null,128]? ? ? 512

_________________________________________________________________

conv_dw_3_relu (ReLU)? ? ? ? [null,null,null,128]? ? ? 0

_________________________________________________________________

conv_pw_3 (Conv2D)? ? ? ? ? ?[null,null,null,128]? ? ? 16384

_________________________________________________________________

conv_pw_3_bn (BatchNormaliza [null,null,null,128]? ? ? 512

_________________________________________________________________

conv_pw_3_relu (ReLU)? ? ? ? [null,null,null,128]? ? ? 0

_________________________________________________________________

conv_pad_4 (ZeroPadding2D)? ?[null,null,null,128]? ? ? 0

_________________________________________________________________

conv_dw_4 (DepthwiseConv2D)? [null,null,null,128]? ? ? 1152

_________________________________________________________________

conv_dw_4_bn (BatchNormaliza [null,null,null,128]? ? ? 512

_________________________________________________________________

conv_dw_4_relu (ReLU)? ? ? ? [null,null,null,128]? ? ? 0

_________________________________________________________________

conv_pw_4 (Conv2D)? ? ? ? ? ?[null,null,null,256]? ? ? 32768

_________________________________________________________________

conv_pw_4_bn (BatchNormaliza [null,null,null,256]? ? ? 1024

_________________________________________________________________

conv_pw_4_relu (ReLU)? ? ? ? [null,null,null,256]? ? ? 0

_________________________________________________________________

conv_dw_5 (DepthwiseConv2D)? [null,null,null,256]? ? ? 2304

_________________________________________________________________

conv_dw_5_bn (BatchNormaliza [null,null,null,256]? ? ? 1024

_________________________________________________________________

conv_dw_5_relu (ReLU)? ? ? ? [null,null,null,256]? ? ? 0

_________________________________________________________________

conv_pw_5 (Conv2D)? ? ? ? ? ?[null,null,null,256]? ? ? 65536

_________________________________________________________________

conv_pw_5_bn (BatchNormaliza [null,null,null,256]? ? ? 1024

_________________________________________________________________

conv_pw_5_relu (ReLU)? ? ? ? [null,null,null,256]? ? ? 0

_________________________________________________________________

conv_pad_6 (ZeroPadding2D)? ?[null,null,null,256]? ? ? 0

_________________________________________________________________

conv_dw_6 (DepthwiseConv2D)? [null,null,null,256]? ? ? 2304

_________________________________________________________________

conv_dw_6_bn (BatchNormaliza [null,null,null,256]? ? ? 1024

_________________________________________________________________

conv_dw_6_relu (ReLU)? ? ? ? [null,null,null,256]? ? ? 0

_________________________________________________________________

conv_pw_6 (Conv2D)? ? ? ? ? ?[null,null,null,512]? ? ? 131072

_________________________________________________________________

conv_pw_6_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_pw_6_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_dw_7 (DepthwiseConv2D)? [null,null,null,512]? ? ? 4608

_________________________________________________________________

conv_dw_7_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_dw_7_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pw_7 (Conv2D)? ? ? ? ? ?[null,null,null,512]? ? ? 262144

_________________________________________________________________

conv_pw_7_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_pw_7_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_dw_8 (DepthwiseConv2D)? [null,null,null,512]? ? ? 4608

_________________________________________________________________

conv_dw_8_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_dw_8_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pw_8 (Conv2D)? ? ? ? ? ?[null,null,null,512]? ? ? 262144

_________________________________________________________________

conv_pw_8_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_pw_8_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_dw_9 (DepthwiseConv2D)? [null,null,null,512]? ? ? 4608

_________________________________________________________________

conv_dw_9_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_dw_9_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pw_9 (Conv2D)? ? ? ? ? ?[null,null,null,512]? ? ? 262144

_________________________________________________________________

conv_pw_9_bn (BatchNormaliza [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_pw_9_relu (ReLU)? ? ? ? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_dw_10 (DepthwiseConv2D) [null,null,null,512]? ? ? 4608

_________________________________________________________________

conv_dw_10_bn (BatchNormaliz [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_dw_10_relu (ReLU)? ? ? ?[null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pw_10 (Conv2D)? ? ? ? ? [null,null,null,512]? ? ? 262144

_________________________________________________________________

conv_pw_10_bn (BatchNormaliz [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_pw_10_relu (ReLU)? ? ? ?[null,null,null,512]? ? ? 0

_________________________________________________________________

conv_dw_11 (DepthwiseConv2D) [null,null,null,512]? ? ? 4608

_________________________________________________________________

conv_dw_11_bn (BatchNormaliz [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_dw_11_relu (ReLU)? ? ? ?[null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pw_11 (Conv2D)? ? ? ? ? [null,null,null,512]? ? ? 262144

_________________________________________________________________

conv_pw_11_bn (BatchNormaliz [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_pw_11_relu (ReLU)? ? ? ?[null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pad_12 (ZeroPadding2D)? [null,null,null,512]? ? ? 0

_________________________________________________________________

conv_dw_12 (DepthwiseConv2D) [null,null,null,512]? ? ? 4608

_________________________________________________________________

conv_dw_12_bn (BatchNormaliz [null,null,null,512]? ? ? 2048

_________________________________________________________________

conv_dw_12_relu (ReLU)? ? ? ?[null,null,null,512]? ? ? 0

_________________________________________________________________

conv_pw_12 (Conv2D)? ? ? ? ? [null,null,null,1024]? ? ?524288

_________________________________________________________________

conv_pw_12_bn (BatchNormaliz [null,null,null,1024]? ? ?4096

_________________________________________________________________

conv_pw_12_relu (ReLU)? ? ? ?[null,null,null,1024]? ? ?0

_________________________________________________________________

conv_dw_13 (DepthwiseConv2D) [null,null,null,1024]? ? ?9216

_________________________________________________________________

conv_dw_13_bn (BatchNormaliz [null,null,null,1024]? ? ?4096

_________________________________________________________________

conv_dw_13_relu (ReLU)? ? ? ?[null,null,null,1024]? ? ?0

_________________________________________________________________

conv_pw_13 (Conv2D)? ? ? ? ? [null,null,null,1024]? ? ?1048576

_________________________________________________________________

conv_pw_13_bn (BatchNormaliz [null,null,null,1024]? ? ?4096

_________________________________________________________________

conv_pw_13_relu (ReLU)? ? ? ?[null,null,null,1024]? ? ?0

=================================================================

Total params: 3228864

Trainable params: 3206976

Non-trainable params: 21888

拿到的模型就只有86層

const?model?=?tf.sequential()

????

????for(let?i?=?0;?i?<?86;?i++)?{

????????const?layer?=?mobilenet.layers[i]

????????layer.trainable?=?false

????????model.add(layer)

????}

????//?連接自己的雙層神經(jīng)網(wǎng)絡(luò)

????model.add(tf.layers.flatten())

調(diào)用model.add(tf.layers.flatten()) 出現(xiàn)下面的報(bào)錯(cuò)

Error: The shape of the input to "Flatten" is not fully defined (got ,,1024). Make sure to pass a complete "input_shape" or "batch_input_shape" argument to the first layer in your model.