当我一步一步运行SSD_training.ipynb文件训练模型的时候，当运行到：
model = SSD300(input_shape, num_classes=NUM_CLASSES)这一步的时候，总是会出现：
File "D:\anaconda\envs\kerasSSD\lib\site-packages\tensorflow\python\framework\common_shapes.py", line 659, in _call_cpp_shape_fn_impl
raise ValueError(err.message)
ValueError: Dimensions must be equal, but are 512 and 38 for 'conv4_3_norm_1/mul' (op: 'Mul') with input shapes: [?,38,38,512], [38].这样的错误
原先我以为是keras、tensorflow版本的问题，后来版本改到您要求的版本后还是会出现
实在不知道怎么解决了，可否指导一下，谢谢！

错误提示：ValueError: Dimensions must be equal, but are 512 and 38 for 'conv4_3_norm/mul' (op: 'Mul') with input shapes: [?,38,38,512], [38].

错误代码：net['conv4_3_norm'] = Normalize(20, name='conv4_3_norm')(net['conv4_3'])

希望能帮忙看一下是哪里出了问题？

C:\Users\wangkai\Desktop\kk344>python get_data_from_XML.py
Traceback (most recent call last):
File "get_data_from_XML.py", line 68, in
data = XML_preprocessor(sys.argv[0]).data
File "get_data_from_XML.py", line 13, in init
self._preprocess_XML()
File "get_data_from_XML.py", line 16, in _preprocess_XML
filenames = os.listdir(self.path_prefix)
NotADirectoryError: [WinError 267] 目录名称无效。: 'get_data_from_XML.py'

你好，在运行你的代码的时候，报了这么一个错误：RuntimeError: module compiled against API version 0xc but this version of numpy is 0xb，网上找了一下原因大概都是说numpy的版本有问题之类的，但是我的环境配置都是跟您readme里面的是一样的，不同的是，由于我是在Windows下跑这个代码的，而tensorflow在windows下只能支持python3，所以我的python版本是3.7的。

Todo: 增加 keras2.+ 版本支持

@kuhung 您好！
请问scikit-video的版本是多少呢，我安装最新的1.1.11出现很多奇怪的问题。谢谢！

你好，我在训练数据时发现GPU的利用率不超过40%，请问能在哪里设置让他的可利用率高一点，以加快训练吗? 十分感谢

环境版本与要求一致，运行python videotest_example.py hy.mp4，提示：
Using TensorFlow backend.
../ssd.py:40: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(64, (3, 3), activation="relu", name="conv1_1", padding="same")
name='conv1_1')(net['input'])
../ssd.py:44: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(64, (3, 3), activation="relu", name="conv1_2", padding="same")
name='conv1_2')(net['conv1_1'])
../ssd.py:46: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool1", padding="same")
name='pool1')(net['conv1_2'])
../ssd.py:51: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (3, 3), activation="relu", name="conv2_1", padding="same")
name='conv2_1')(net['pool1'])
../ssd.py:55: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (3, 3), activation="relu", name="conv2_2", padding="same")
name='conv2_2')(net['conv2_1'])
../ssd.py:57: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool2", padding="same")
name='pool2')(net['conv2_2'])
../ssd.py:62: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv3_1", padding="same")
name='conv3_1')(net['pool2'])
../ssd.py:66: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv3_2", padding="same")
name='conv3_2')(net['conv3_1'])
../ssd.py:70: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv3_3", padding="same")
name='conv3_3')(net['conv3_2'])
../ssd.py:72: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool3", padding="same")
name='pool3')(net['conv3_3'])
../ssd.py:77: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv4_1", padding="same")
name='conv4_1')(net['pool3'])
../ssd.py:81: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv4_2", padding="same")
name='conv4_2')(net['conv4_1'])
../ssd.py:85: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv4_3", padding="same")
name='conv4_3')(net['conv4_2'])
../ssd.py:87: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool4", padding="same")
name='pool4')(net['conv4_3'])
../ssd.py:92: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv5_1", padding="same")
name='conv5_1')(net['pool4'])
../ssd.py:96: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv5_2", padding="same")
name='conv5_2')(net['conv5_1'])
../ssd.py:100: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv5_3", padding="same")
name='conv5_3')(net['conv5_2'])
../ssd.py:102: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((3, 3), strides=(1, 1), name="pool5", padding="same")
name='pool5')(net['conv5_3'])
/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/legacy/layers.py:301: UserWarning: The AtrousConvolution2D layer has been deprecated. Use instead the Conv2D layer with the dilation_rate argument.
warnings.warn('The AtrousConvolution2D layer '
/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/legacy/layers.py:305: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(1024, (3, 3), activation="relu", name="fc6", dilation_rate=(6, 6), padding="same")
return Conv2D(*args, **kwargs)
../ssd.py:110: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(1024, (1, 1), activation="relu", name="fc7", padding="same")
border_mode='same', name='fc7')(net['fc6'])
../ssd.py:115: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (1, 1), activation="relu", name="conv6_1", padding="same")
name='conv6_1')(net['fc7'])
../ssd.py:118: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv6_2", strides=(2, 2), padding="same")
name='conv6_2')(net['conv6_1'])
../ssd.py:122: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (1, 1), activation="relu", name="conv7_1", padding="same")
name='conv7_1')(net['conv6_2'])
../ssd.py:126: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv7_2", strides=(2, 2), padding="valid")
name='conv7_2')(net['conv7_2'])
../ssd.py:130: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (1, 1), activation="relu", name="conv8_1", padding="same")
name='conv8_1')(net['conv7_2'])
../ssd.py:133: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv8_2", strides=(2, 2), padding="same")
name='conv8_2')(net['conv8_1'])
../ssd.py:140: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(12, (3, 3), name="conv4_3_norm_mbox_loc", padding="same")
name='conv4_3_norm_mbox_loc')(net['conv4_3_norm'])
../ssd.py:148: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(63, (3, 3), name="conv4_3_norm_mbox_conf", padding="same")
name=name)(net['conv4_3_norm'])
../ssd.py:160: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="fc7_mbox_loc", padding="same")
name='fc7_mbox_loc')(net['fc7'])
../ssd.py:168: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="fc7_mbox_conf", padding="same")
name=name)(net['fc7'])
../ssd.py:178: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="conv6_2_mbox_loc", padding="same")
name='conv6_2_mbox_loc')(net['conv6_2'])
../ssd.py:186: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="conv6_2_mbox_conf", padding="same")
name=name)(net['conv6_2'])
../ssd.py:197: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="conv7_2_mbox_loc", padding="same")
name='conv7_2_mbox_loc')(net['conv7_2'])
../ssd.py:205: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="conv7_2_mbox_conf", padding="same")
name=name)(net['conv7_2'])
../ssd.py:216: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="conv8_2_mbox_loc", padding="same")
name='conv8_2_mbox_loc')(net['conv8_2'])
../ssd.py:224: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="conv8_2_mbox_conf", padding="same")
name=name)(net['conv8_2'])
Traceback (most recent call last):
File "videotest_example.py", line 16, in
model = SSD(input_shape, num_classes=NUM_CLASSES)
File "../ssd.py", line 258, in SSD300
mode='concat', concat_axis=1, name='mbox_loc')
TypeError: 'module' object is not callable
上网查了之后，将ssd.py中修改如下：
net['mbox_loc'] = concatenate([net['conv4_3_norm_mbox_loc_flat'],
net['fc7_mbox_loc_flat'],
net['conv6_2_mbox_loc_flat'],
net['conv7_2_mbox_loc_flat'],
net['conv8_2_mbox_loc_flat'],
net['pool6_mbox_loc_flat']],
axis=1, name='mbox_loc')
net['mbox_conf'] = concatenate([net['conv4_3_norm_mbox_conf_flat'],
net['fc7_mbox_conf_flat'],
net['conv6_2_mbox_conf_flat'],
net['conv7_2_mbox_conf_flat'],
net['conv8_2_mbox_conf_flat'],
net['pool6_mbox_conf_flat']],
axis=1, name='mbox_conf')
net['mbox_priorbox'] = concatenate([net['conv4_3_norm_mbox_priorbox'],
net['fc7_mbox_priorbox'],
net['conv6_2_mbox_priorbox'],
net['conv7_2_mbox_priorbox'],
net['conv8_2_mbox_priorbox'],
net['pool6_mbox_priorbox']],
axis=1,
name='mbox_priorbox')
之后报错如下：
Using TensorFlow backend.
../ssd.py:41: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(64, (3, 3), activation="relu", name="conv1_1", padding="same")
name='conv1_1')(net['input'])
../ssd.py:45: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(64, (3, 3), activation="relu", name="conv1_2", padding="same")
name='conv1_2')(net['conv1_1'])
../ssd.py:47: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool1", padding="same")
name='pool1')(net['conv1_2'])
../ssd.py:52: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (3, 3), activation="relu", name="conv2_1", padding="same")
name='conv2_1')(net['pool1'])
../ssd.py:56: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (3, 3), activation="relu", name="conv2_2", padding="same")
name='conv2_2')(net['conv2_1'])
../ssd.py:58: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool2", padding="same")
name='pool2')(net['conv2_2'])
../ssd.py:63: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv3_1", padding="same")
name='conv3_1')(net['pool2'])
../ssd.py:67: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv3_2", padding="same")
name='conv3_2')(net['conv3_1'])
../ssd.py:71: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv3_3", padding="same")
name='conv3_3')(net['conv3_2'])
../ssd.py:73: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool3", padding="same")
name='pool3')(net['conv3_3'])
../ssd.py:78: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv4_1", padding="same")
name='conv4_1')(net['pool3'])
../ssd.py:82: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv4_2", padding="same")
name='conv4_2')(net['conv4_1'])
../ssd.py:86: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv4_3", padding="same")
name='conv4_3')(net['conv4_2'])
../ssd.py:88: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((2, 2), strides=(2, 2), name="pool4", padding="same")
name='pool4')(net['conv4_3'])
../ssd.py:93: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv5_1", padding="same")
name='conv5_1')(net['pool4'])
../ssd.py:97: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv5_2", padding="same")
name='conv5_2')(net['conv5_1'])
../ssd.py:101: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv5_3", padding="same")
name='conv5_3')(net['conv5_2'])
../ssd.py:103: UserWarning: Update your MaxPooling2D call to the Keras 2 API: MaxPooling2D((3, 3), strides=(1, 1), name="pool5", padding="same")
name='pool5')(net['conv5_3'])
/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/legacy/layers.py:301: UserWarning: The AtrousConvolution2D layer has been deprecated. Use instead the Conv2D layer with the dilation_rate argument.
warnings.warn('The AtrousConvolution2D layer '
/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/legacy/layers.py:305: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(1024, (3, 3), activation="relu", name="fc6", dilation_rate=(6, 6), padding="same")
return Conv2D(*args, **kwargs)
../ssd.py:111: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(1024, (1, 1), activation="relu", name="fc7", padding="same")
border_mode='same', name='fc7')(net['fc6'])
../ssd.py:116: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (1, 1), activation="relu", name="conv6_1", padding="same")
name='conv6_1')(net['fc7'])
../ssd.py:119: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(512, (3, 3), activation="relu", name="conv6_2", strides=(2, 2), padding="same")
name='conv6_2')(net['conv6_1'])
../ssd.py:123: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (1, 1), activation="relu", name="conv7_1", padding="same")
name='conv7_1')(net['conv6_2'])
../ssd.py:127: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv7_2", strides=(2, 2), padding="valid")
name='conv7_2')(net['conv7_2'])
../ssd.py:131: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(128, (1, 1), activation="relu", name="conv8_1", padding="same")
name='conv8_1')(net['conv7_2'])
../ssd.py:134: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(256, (3, 3), activation="relu", name="conv8_2", strides=(2, 2), padding="same")
name='conv8_2')(net['conv8_1'])
../ssd.py:141: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(12, (3, 3), name="conv4_3_norm_mbox_loc", padding="same")
name='conv4_3_norm_mbox_loc')(net['conv4_3_norm'])
../ssd.py:149: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(63, (3, 3), name="conv4_3_norm_mbox_conf", padding="same")
name=name)(net['conv4_3_norm'])
../ssd.py:161: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="fc7_mbox_loc", padding="same")
name='fc7_mbox_loc')(net['fc7'])
../ssd.py:169: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="fc7_mbox_conf", padding="same")
name=name)(net['fc7'])
../ssd.py:179: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="conv6_2_mbox_loc", padding="same")
name='conv6_2_mbox_loc')(net['conv6_2'])
../ssd.py:187: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="conv6_2_mbox_conf", padding="same")
name=name)(net['conv6_2'])
../ssd.py:198: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="conv7_2_mbox_loc", padding="same")
name='conv7_2_mbox_loc')(net['conv7_2'])
../ssd.py:206: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="conv7_2_mbox_conf", padding="same")
name=name)(net['conv7_2'])
../ssd.py:217: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(24, (3, 3), name="conv8_2_mbox_loc", padding="same")
name='conv8_2_mbox_loc')(net['conv8_2'])
../ssd.py:225: UserWarning: Update your Conv2D call to the Keras 2 API: Conv2D(126, (3, 3), name="conv8_2_mbox_conf", padding="same")
name=name)(net['conv8_2'])
Traceback (most recent call last):
File "videotest_example.py", line 16, in
model = SSD(input_shape, num_classes=NUM_CLASSES)
File "../ssd.py", line 274, in SSD300
name='mbox_priorbox')
File "/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/layers/merge.py", line 641, in concatenate
return Concatenate(axis=axis, **kwargs)(inputs)
File "/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/engine/base_layer.py", line 434, in call
self.build(input_shapes)
File "/home/zwc/anaconda3/envs/ssd-video/lib/python3.6/site-packages/keras/layers/merge.py", line 354, in build
'Got inputs shapes: %s' % (input_shape))
ValueError: A Concatenate layer requires inputs with matching shapes except for the concat axis. Got inputs shapes: [(None, 38, 38, 512), (None, 19, 19, 1024), (None, 10, 10, 512), (None, 5, 5, 256), (None, 3, 3, 256), (None, 1, 1, 256)]

您好！
我现在已经可以按照交通标志识别的步骤可以训练出自己的模型了(教程中的三种)
我想要增大训练样本和种类，比如进行四种交通标志的识别，请问
1、预训练权重weights_SSD300.hdf5是否可以继续使用？
2、最后预测的结果图片上预测种类为何显示为“1”“2”“3”的数字？如何修改为种类名称，比如“packing”之类的名称。
感谢回复

已经做成了VOC格式，应该怎么训练？

@kuhung
您好！
请问，prior_boxes_ssd300.pkl 这个文件是怎么得到的，谢谢！

ValueError Traceback (most recent call last)
in
5 validation_data=gen.generate(False),
6 validation_steps=num_val,
----> 7 nb_worker=0)

/opt/conda/lib/python3.6/site-packages/keras/legacy/interfaces.py in wrapper(*args, **kwargs)
89 warnings.warn('Update your ' + object_name + ' call to the ' +
90 'Keras 2 API: ' + signature, stacklevel=2)
---> 91 return func(*args, **kwargs)
92 wrapper._original_function = func
93 return wrapper

/opt/conda/lib/python3.6/site-packages/keras/engine/training.py in fit_generator(self, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
1416 use_multiprocessing=use_multiprocessing,
1417 shuffle=shuffle,
-> 1418 initial_epoch=initial_epoch)
1419
1420 @interfaces.legacy_generator_methods_support

/opt/conda/lib/python3.6/site-packages/keras/engine/training_generator.py in fit_generator(model, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
179 batch_index = 0
180 while steps_done < steps_per_epoch:
--> 181 generator_output = next(output_generator)
182
183 if not hasattr(generator_output, 'len'):

in generate(self, train)
157 if self.vflip_prob > 0:
158 img, y = self.vertical_flip(img, y)
--> 159 y = self.bbox_util.assign_boxes(y)
160 inputs.append(img)
161 targets.append(y)

/kaggle/working/ssd_utils.py in assign_boxes(self, boxes)
147 :4]
148 assignment[:, 4][best_iou_mask] = 0
--> 149 assignment[:, 5:-8][best_iou_mask] = boxes[best_iou_idx, 4:]
150 assignment[:, -8][best_iou_mask] = 1
151 return assignment

ValueError: shape mismatch: value array of shape (4,2) could not be broadcast to indexing result of shape (4,1)

但我在 jupyter中运行SSD.ipynb 是出现错误，不知道如何解决。代码如下：
input_shape=(300, 300, 3)
model = SSD300(input_shape, num_classes=NUM_CLASSES)
model.load_weights('weights.04-2.13.hdf5', by_name=True)
bbox_util = BBoxUtility(NUM_CLASSES)

错误信息如下：

TypeError Traceback (most recent call last)
in
1 input_shape=(300, 300, 3)
----> 2 model = SSD300(input_shape, num_classes=NUM_CLASSES)
3 model.load_weights('/home/syk/jupyter/SSD_keras/weights.04-2.13.hdf5', by_name=True)
4 bbox_util = BBoxUtility(NUM_CLASSES)

~/jupyter/SSD_keras/ssd.py in SSD300(input_shape, num_classes)
256 net['conv8_2_mbox_loc_flat'],
257 net['pool6_mbox_loc_flat']],
--> 258 mode='concat', concat_axis=1, name='mbox_loc')
259 net['mbox_conf'] = merge([net['conv4_3_norm_mbox_conf_flat'],
260 net['fc7_mbox_conf_flat'],

TypeError: 'module' object is not callable

File "C:\Users\Administrator\Desktop\SSD_keras-master\SSD_keras-master\ssd_layers.py", line 47, in call
output = K.l2_normalize(x, 3)

File "C:\Program Files\Anaconda3\lib\site-packages\keras\backend\tensorflow_backend.py", line 3157, in l2_normalize

TypeError: l2_normalize() got an unexpected keyword argument 'axis'

skvideo.io.VideoCapture can't be used