|
创建自定义的SSD模型结构(基于keras)
Model: "SSD_dfy_training"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) (None, 300, 480, 3) 0
__________________________________________________________________________________________________
identity_layer (Lambda) (None, 300, 480, 3) 0 input_1[0][0]
__________________________________________________________________________________________________
input_subtract_mean (Lambda) (None, 300, 480, 3) 0 identity_layer[0][0]
__________________________________________________________________________________________________
input_divide_by_stddev (Lambda) (None, 300, 480, 3) 0 input_subtract_mean[0][0]
__________________________________________________________________________________________________
conv1 (Conv2D) (None, 300, 480, 32) 2432 input_divide_by_stddev[0][0]
__________________________________________________________________________________________________
bn1 (BatchNormalization) (None, 300, 480, 32) 128 conv1[0][0]
__________________________________________________________________________________________________
elu1 (ELU) (None, 300, 480, 32) 0 bn1[0][0]
__________________________________________________________________________________________________
pool1 (MaxPooling2D) (None, 150, 240, 32) 0 elu1[0][0]
__________________________________________________________________________________________________
conv2 (Conv2D) (None, 150, 240, 48) 13872 pool1[0][0]
__________________________________________________________________________________________________
bn2 (BatchNormalization) (None, 150, 240, 48) 192 conv2[0][0]
__________________________________________________________________________________________________
elu2 (ELU) (None, 150, 240, 48) 0 bn2[0][0]
__________________________________________________________________________________________________
pool2 (MaxPooling2D) (None, 75, 120, 48) 0 elu2[0][0]
__________________________________________________________________________________________________
conv3 (Conv2D) (None, 75, 120, 64) 27712 pool2[0][0]
__________________________________________________________________________________________________
bn3 (BatchNormalization) (None, 75, 120, 64) 256 conv3[0][0]
__________________________________________________________________________________________________
elu3 (ELU) (None, 75, 120, 64) 0 bn3[0][0]
__________________________________________________________________________________________________
pool3 (MaxPooling2D) (None, 37, 60, 64) 0 elu3[0][0]
__________________________________________________________________________________________________
conv4 (Conv2D) (None, 37, 60, 64) 36928 pool3[0][0]
__________________________________________________________________________________________________
bn4 (BatchNormalization) (None, 37, 60, 64) 256 conv4[0][0]
__________________________________________________________________________________________________
elu4 (ELU) (None, 37, 60, 64) 0 bn4[0][0]
__________________________________________________________________________________________________
pool4 (MaxPooling2D) (None, 18, 30, 64) 0 elu4[0][0]
__________________________________________________________________________________________________
conv5 (Conv2D) (None, 18, 30, 48) 27696 pool4[0][0]
__________________________________________________________________________________________________
bn5 (BatchNormalization) (None, 18, 30, 48) 192 conv5[0][0]
__________________________________________________________________________________________________
elu5 (ELU) (None, 18, 30, 48) 0 bn5[0][0]
__________________________________________________________________________________________________
pool5 (MaxPooling2D) (None, 9, 15, 48) 0 elu5[0][0]
__________________________________________________________________________________________________
conv6 (Conv2D) (None, 9, 15, 48) 20784 pool5[0][0]
__________________________________________________________________________________________________
bn6 (BatchNormalization) (None, 9, 15, 48) 192 conv6[0][0]
__________________________________________________________________________________________________
elu6 (ELU) (None, 9, 15, 48) 0 bn6[0][0]
__________________________________________________________________________________________________
pool6 (MaxPooling2D) (None, 4, 7, 48) 0 elu6[0][0]
__________________________________________________________________________________________________
conv7 (Conv2D) (None, 4, 7, 32) 13856 pool6[0][0]
__________________________________________________________________________________________________
bn7 (BatchNormalization) (None, 4, 7, 32) 128 conv7[0][0]
__________________________________________________________________________________________________
elu7 (ELU) (None, 4, 7, 32) 0 bn7[0][0]
__________________________________________________________________________________________________
classes4 (Conv2D) (None, 37, 60, 24) 13848 elu4[0][0]
__________________________________________________________________________________________________
classes5 (Conv2D) (None, 18, 30, 24) 10392 elu5[0][0]
__________________________________________________________________________________________________
classes6 (Conv2D) (None, 9, 15, 24) 10392 elu6[0][0]
__________________________________________________________________________________________________
classes7 (Conv2D) (None, 4, 7, 24) 6936 elu7[0][0]
__________________________________________________________________________________________________
boxes4 (Conv2D) (None, 37, 60, 16) 9232 elu4[0][0]
__________________________________________________________________________________________________
boxes5 (Conv2D) (None, 18, 30, 16) 6928 elu5[0][0]
__________________________________________________________________________________________________
boxes6 (Conv2D) (None, 9, 15, 16) 6928 elu6[0][0]
__________________________________________________________________________________________________
boxes7 (Conv2D) (None, 4, 7, 16) 4624 elu7[0][0]
__________________________________________________________________________________________________
classes4_reshaped (Reshape) (None, 8880, 6) 0 classes4[0][0]
__________________________________________________________________________________________________
classes5_reshaped (Reshape) (None, 2160, 6) 0 classes5[0][0]
__________________________________________________________________________________________________
classes6_reshaped (Reshape) (None, 540, 6) 0 classes6[0][0]
__________________________________________________________________________________________________
classes7_reshaped (Reshape) (None, 112, 6) 0 classes7[0][0]
__________________________________________________________________________________________________
anchors4 (AnchorBoxes) (None, 37, 60, 4, 8) 0 boxes4[0][0]
__________________________________________________________________________________________________
anchors5 (AnchorBoxes) (None, 18, 30, 4, 8) 0 boxes5[0][0]
__________________________________________________________________________________________________
anchors6 (AnchorBoxes) (None, 9, 15, 4, 8) 0 boxes6[0][0]
__________________________________________________________________________________________________
anchors7 (AnchorBoxes) (None, 4, 7, 4, 8) 0 boxes7[0][0]
__________________________________________________________________________________________________
classes_concat (Concatenate) (None, 11692, 6) 0 classes4_reshaped[0][0]
classes5_reshaped[0][0]
classes6_reshaped[0][0]
classes7_reshaped[0][0]
__________________________________________________________________________________________________
boxes4_reshaped (Reshape) (None, 8880, 4) 0 boxes4[0][0]
__________________________________________________________________________________________________
boxes5_reshaped (Reshape) (None, 2160, 4) 0 boxes5[0][0]
__________________________________________________________________________________________________
boxes6_reshaped (Reshape) (None, 540, 4) 0 boxes6[0][0]
__________________________________________________________________________________________________
boxes7_reshaped (Reshape) (None, 112, 4) 0 boxes7[0][0]
__________________________________________________________________________________________________
anchors4_reshaped (Reshape) (None, 8880, 8) 0 anchors4[0][0]
__________________________________________________________________________________________________
anchors5_reshaped (Reshape) (None, 2160, 8) 0 anchors5[0][0]
__________________________________________________________________________________________________
anchors6_reshaped (Reshape) (None, 540, 8) 0 anchors6[0][0]
__________________________________________________________________________________________________
anchors7_reshaped (Reshape) (None, 112, 8) 0 anchors7[0][0]
__________________________________________________________________________________________________
classes_concat_softmax (Activat (None, 11692, 6) 0 classes_concat[0][0]
__________________________________________________________________________________________________
boxes_concat (Concatenate) (None, 11692, 4) 0 boxes4_reshaped[0][0]
boxes5_reshaped[0][0]
boxes6_reshaped[0][0]
boxes7_reshaped[0][0]
__________________________________________________________________________________________________
anchors_concat (Concatenate) (None, 11692, 8) 0 anchors4_reshaped[0][0]
anchors5_reshaped[0][0]
anchors6_reshaped[0][0]
anchors7_reshaped[0][0]
__________________________________________________________________________________________________
predictions (Concatenate) (None, 11692, 18) 0 classes_concat_softmax[0][0]
boxes_concat[0][0]
anchors_concat[0][0]
==================================================================================================
Total params: 213,904
Trainable params: 213,232
Non-trainable params: 672
__________________________________________________________________________________________________
Process finished with exit code 0
|
|