Tensorflow:如何将自定义输入插入现有图表?

时间:2016-07-27 16:50:48

标签: graph tensorflow subgraph

我已经下载了一个实现VGG16 ConvNet的张量流GraphDef,我用它做了这个:

Pl['images'] = tf.placeholder(tf.float32, 
                          [None, 448, 448, 3],
                          name="images") #batch x width x height x channels
with open("tensorflow-vgg16/vgg16.tfmodel", mode='rb') as f: 
    fileContent = f.read()

graph_def = tf.GraphDef()
graph_def.ParseFromString(fileContent)
tf.import_graph_def(graph_def, input_map={"images": Pl['images']})

此外,我的图像特征与"import/pool5/"的输出是同质的。

如何告诉我的图表不想使用他的输入"images",而是将张量"import/pool5/"作为输入?

谢谢!

修改

好的,我意识到我不是很清楚。情况如下:

我正在尝试使用this implementation的ROI池,使用预先训练的VGG16,我有GraphDef格式。所以这就是我的工作:

首先,我加载模型:

tf.reset_default_graph()
with open("tensorflow-vgg16/vgg16.tfmodel",
          mode='rb') as f:
    fileContent = f.read()
graph_def = tf.GraphDef()
graph_def.ParseFromString(fileContent)
graph = tf.get_default_graph()

然后,我创建了占位符

images = tf.placeholder(tf.float32, 
                              [None, 448, 448, 3],
                              name="images") #batch x width x height x channels
boxes = tf.placeholder(tf.float32, 
                             [None,5], # 5 = [batch_id,x1,y1,x2,y2]
                             name = "boxes")

我将图的第一部分的输出定义为conv5_3 / Relu

tf.import_graph_def(graph_def, 
                    input_map={'images':images})
out_tensor = graph.get_tensor_by_name("import/conv5_3/Relu:0")

因此,out_tensor的形状为[None,14,14,512]

然后,我进行投资回报汇集:

[out_pool,argmax] = module.roi_pool(out_tensor,
                                    boxes,
                                    7,7,1.0/1) 

使用out_pool.shape = N_Boxes_in_batch x 7 x 7 x 512,与pool5同质。然后,我想将out_pool作为pool5之后的操作的输入,因此它看起来像

tf.import_graph_def(graph.as_graph_def(),
                    input_map={'import/pool5':out_pool})

但它不起作用,我有这个错误:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-89-527398d7344b> in <module>()
      5 
      6 tf.import_graph_def(graph.as_graph_def(),
----> 7                     input_map={'import/pool5':out_pool})
      8 
      9 final_out = graph.get_tensor_by_name("import/Relu_1:0")

/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/importer.py in import_graph_def(graph_def, input_map, return_elements, name, op_dict)
    333       # NOTE(mrry): If the graph contains a cycle, the full shape information
    334       # may not be available for this op's inputs.
--> 335       ops.set_shapes_for_outputs(op)
    336 
    337       # Apply device functions for this op.

/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/ops.py in set_shapes_for_outputs(op)
   1610       raise RuntimeError("No shape function registered for standard op: %s"
   1611                          % op.type)
-> 1612   shapes = shape_func(op)
   1613   if len(op.outputs) != len(shapes):
   1614     raise RuntimeError(

/home/hbenyounes/vqa/roi_pooling_op_grad.py in _roi_pool_shape(op)
     13   channels = dims_data[3]
     14   print(op.inputs[1].name, op.inputs[1].get_shape())
---> 15   dims_rois = op.inputs[1].get_shape().as_list()
     16   num_rois = dims_rois[0]
     17 

/usr/local/lib/python3.4/dist-packages/tensorflow/python/framework/tensor_shape.py in as_list(self)
    745       A list of integers or None for each dimension.
    746     """
--> 747     return [dim.value for dim in self._dims]
    748 
    749   def as_proto(self):

TypeError: 'NoneType' object is not iterable

有任何线索吗?

3 个答案:

答案 0 :(得分:1)

我会做的是这样的事情:

- 首先检索张量的名称,这些张量代表VGG16中pool5之后的3个完全连接层的权重和偏差。
要做到这一点,我会检查[n.name for n in graph.as_graph_def().node]。 (它们可能看起来像import / locali / weight:0,import / locali / bias:0等)

- 将它们放在python列表中:

weights_names=["import/local1/weight:0" ,"import/local2/weight:0" ,"import/local3/weight:0"]
biases_names=["import/local1/bias:0" ,"import/local2/bias:0" ,"import/local3/bias:0"]

- 定义一个类似于:

的函数
def pool5_tofcX(input_tensor, layer_number=3):
  flatten=tf.reshape(input_tensor,(-1,7*7*512))
  tmp=flatten
  for i in xrange(layer_number):
    tmp=tf.matmul(tmp, graph.get_tensor_by_name(weights_name[i]))
    tmp=tf.nn.bias_add(tmp, graph.get_tensor_by_name(biases_name[i]))
    tmp=tf.nn.relu(tmp)
  return tmp

然后使用函数定义张量:

wanted_output=pool5_tofcX(out_pool) 

然后你就完成了!

答案 1 :(得分:1)

使用tf.train.export_meta_graph来存储整个MetaGraph通常非常方便。然后,在恢复时,您可以使用tf.train.import_meta_graph,因为结果是它将所有其他参数传递给具有import_scoped_meta_graph参数的基础input_map并使用它当它到达它自己的import_graph_def调用时。

没有记录,并且花了很多时间才找到它,但它确实有效!

答案 2 :(得分:0)

Jonan Georgiev在这里提供了一个很好的答案。在此git问题的结尾,还描述了相同的方法,但几乎没有大张声调:https://github.com/tensorflow/tensorflow/issues/3389

下面是使用此方法为getQueryFromDB(req, res, next){ let query = { accountType: req.params.userType, isAdmin: false }; Profile.find(query) .lean() .then(users => { return res.json(users); }) .catch(err => { console.log(err) }) } tf.data.Dataset张量切换占位符的复制/粘贴可运行示例。

get_next

输出,我们可以清楚地看到平方运算的输入已更改。

import tensorflow as tf


my_placeholder = tf.placeholder(dtype=tf.float32, shape=1, name='my_placeholder')
my_op = tf.square(my_placeholder, name='my_op')

# Save the graph to memory
graph_def = tf.get_default_graph().as_graph_def()

print('----- my_op before any remapping -----')
print([n for n in graph_def.node if n.name == 'my_op'])

tf.reset_default_graph()

ds = tf.data.Dataset.from_tensors(1.0)
next_tensor = tf.data.make_one_shot_iterator(ds).get_next(name='my_next_tensor')

# Restore the graph with a custom input mapping
tf.graph_util.import_graph_def(graph_def, input_map={'my_placeholder': next_tensor}, name='')

print('----- my_op after remapping -----')
print([n for n in tf.get_default_graph().as_graph_def().node if n.name == 'my_op'])