修正Tensorflow中的去卷积层

我试图用Tensorflow构建一个变分自动编码器。我从最简单的模型开始。我有以下方法：修正Tensorflow中的去卷积层

def conv_layer(x, w_shape, b_shape, padding='SAME'): 
    W = weight_variable(w_shape) 
    tf.summary.histogram(W.name, W) 

    b = bias_variable(b_shape) 
    tf.summary.histogram(b.name, b) 

    # Note that I used a stride of 2 on purpose in order not to use max pool layer. 
    activations = tf.nn.relu(tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding=padding) + b) 
    tf.summary.histogram(activations.name, activations) 
    return activations 

def deconv_layer(x, w_shape, b_shape, padding="SAME"): 
    W = weight_variable(w_shape) 
    tf.summary.histogram(W.name, W) 

    b = bias_variable(b_shape) 
    tf.summary.histogram('bias', b) 

    x_shape = tf.shape(x) 

    out_shape = tf.stack([x_shape[0], x_shape[1], x_shape[2], w_shape[2]]) 
    # Note that I have used a stride of 2 since I used a stride of 2 in conv layer. 
    transposed_activations = tf.nn.conv2d_transpose(x, W, out_shape, [1, 1, 1, 1], padding=padding) + b 
    tf.summary.histogram(transposed_activations.name, transposed_activations) 
    return transposed_activations

而整个网络的模型如下：

with tf.name_scope('conv1'): 
    conv1 = conv_layer(image, [3, 3, 3, 32], [32]) 
with tf.name_scope('conv2'): 
    conv2 = conv_layer(conv1, [3, 3, 32, 64], [64]) 
with tf.name_scope('conv3'): 
    conv3 = conv_layer(conv2, [3, 3, 64, 128], [128]) 
with tf.name_scope('conv4'): 
    conv4 = conv_layer(conv3, [3, 3, 128, 256], [256]) 

with tf.name_scope('z'): 
    z = conv_layer(conv4, [3, 3, 256, 256], [256]) 

with tf.name_scope('deconv4'): 
    deconv4 = deconv_layer(z, [3, 3, 128, 256], [128]) 
with tf.name_scope('deconv3'): 
    deconv3 = deconv_layer(deconv4, [3, 3, 64, 128], [64]) 
with tf.name_scope('deconv2'): 
    deconv2 = deconv_layer(deconv3, [3, 3, 32, 64], [32]) 
with tf.name_scope('deconv1'): 
    deconv_image = deconv_layer(deconv2, [3, 3, 3, 32], [3])

我从FIFOQueue让我的图像，并将它们送入这一模式。我的图像尺寸是112, 112, 3。我的问题是在这两个CONV和deconv层我有以下错误从 [1, 1, 1, 1] to [1, 2, 2, 1]改变步幅时：

InvalidArgumentError (see above for traceback): Conv2DSlowBackpropInput: Size of out_backprop doesn't match computed: actual = 4, computed = 2 
    [[Node: deconv4/conv2d_transpose = Conv2DBackpropInput[T=DT_FLOAT, data_format="NHWC", padding="SAME", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true, _device="/job:localhost/replica:0/task:0/gpu:0"](deconv4/stack, deconv4/Variable/read, z/Relu)]] 
    [[Node: deconv1/add/_17 = _Recv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_85_deconv1/add", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/cpu:0"]()]]

PS：我知道，我的思念在deconv层的激活功能，但我猜这与我得到的错误没有关系。任何帮助非常感谢！

来源

2017-07-07 I. A

我有同样的错误。你解决了吗？ – freude

@幸运，希望我的回答对你有意义。请喜欢，如果你确信。 –

@freude，我很好地调整了解决方案。我忘了提及输出形状的问题。请检讨最终答案，并接受它，如果你确信！谢谢 –

这样做的原因错误是如下：

如果我们假定我们有stride = 2在每个CONV层，然后，在我的情况下，当输入图像的大小为112, 112, 3，每个conv layer大小后frames或feature maps在应用卷积后减半。也就是说，在conv1之后，图像的大小（高度，宽度）变为[56, 56]。在conv2之后，尺寸变成[28, 28]。 conv3之后：[14, 14]，之后conv4：[7, 7]。因此，应用一个名为z的额外conv layer已将尺寸减小到[3, 3]。这是问题：7不能被2整除。所以我们得到了不同的维度。从[3, 3]到[112, 112]从申请deconv layer后是不可能的。此外：

[3, 3] -> [6, 6] -> [12, 12] -> [24, 24] -> [48, 48]

第二错误：在deconv layer输出形状应该如下：

# we should multiply x_shape[1] and x_shape[2] by 2. 
out_shape = tf.stack([x_shape[0], x_shape[1] * 2, x_shape[2] * 2, w_shape[2]])

因此，最终deconv layer变为如下：

def deconv_layer(x, w_shape, b_shape, is_training, padding="SAME", activation='selu'): 
    W = weight_variable(w_shape) 
    tf.summary.histogram("weights", W) 

    b = bias_variable(b_shape) 
    tf.summary.histogram('biases', b) 

    x_shape = tf.shape(x) 
    # output shape: [batch_size, h * 2, w * 2, input_shape from w]. 
    out_shape = tf.stack([x_shape[0], x_shape[1] * 2, x_shape[2] * 2, w_shape[2]]) 
    # Note that I have used a stride of 2 since I used a stride of 2 in conv layer. 
    if activation == 'selu': 
     conv_trans = tf.nn.conv2d_transpose(x, W, out_shape, [1, 2, 2, 1], padding=padding) + b 

     transposed_activations = tf.nn.elu(conv_trans) 

    else: 
     conv_trans = tf.nn.conv2d_transpose(x, W, out_shape, [1, 2, 2, 1], padding=padding) + b 
     transposed_activations = tf.nn.sigmoid(conv_trans) 

    tf.summary.histogram("transpose_activation", transposed_activations) 
    return transposed_activations

因此，大小的输出与输入的大小不同，这就是我得到错误的原因。而为了做back propagation我们需要一个成本函数。此成本函数将取决于output和input。因此，如果他们有不同的大小，那会导致错误。

可以解决的方法是让conv layer, z,拥有stride of 1。

来源

2017-08-02 22:17:17

当上述代码中的步幅为[1, 2, 2, 1]时，input和deconv_image的形状不匹配。这种情况下的decov_image大小将为[64x64]。你在代码中的其他地方照顾这个吗？

来源

2017-07-08 05:36:42

当我修改了deconv层的out_shape到： 'out_shape = tf.stack（[x_shape [0]，x_shape [1] * 2，x_shape [2] * 2，w_shape [2]]）'It解决了。也许这是错误的根源。这是有道理的。 –

修正Tensorflow中的去卷积层

回答

相关问题