Wrapper for :class:`Conv2dLayer`, if you don't understand how to use :class:`Conv2dLayer`, this function may be easier. Parameters ----------

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

:param path: the input file path :param image_size: the input image size :return: the returned datasets is opencv format BGR [112, 112,

:param embedding: the input embedding vectors :param labels: the input labels, the shape should be eg: (batch_size, 1) :param s: scalar

Reference slim resnet :param inputs: :param num_outputs: :param kernel_size: :param strides: :param rate: :param scope:

Reference slim resnet :param inputs: :param num_outputs: :param kernel_size: :param strides: :param rate: :param scope:

:param inputs: inputs is an tensorflow placeholder :param num_classes: :param num_layers: :param type: choose weather using se xt or

return data and label on ctx

Reference slim resnet :param inputs: :param num_outputs: :param kernel_size: :param strides: :param rate: :param scope:

Reference slim resnet :param inputs: :param num_outputs: :param kernel_size: :param strides: :param rate: :param scope:

Reference slim resnet :param inputs: :param num_outputs: :param kernel_size: :param strides: :param rate: :param scope:

Build the VGG 19 Model Parameters ----------- rgb : rgb image placeholder [batch, height, width, 3] values scaled [0, 1]

:param embedding: the input embedding vectors :param labels: the input labels, the shape should be eg: (batch_size, 1) :param s: scalar

Calculate the average gradient for each shared variable across all towers. Note that this function provides a synchronization point across all towe

Calculate the average gradient for each shared variable across all towers. Note that this function provides a synchronization point across all towe

Calculate the average gradient for each shared variable across all towers. Note that this function provides a synchronization point across all towe

Copy from [insightface](https://github.com/deepinsight/insightface) :param thresholds: :param embeddings1: :param embeddings2: :p

Reference slim resnet :param inputs: :param num_outputs: :param kernel_size: :param strides: :param rate: :param scope:

:param embedding: the input embedding vectors :param labels: the input labels, the shape should be eg: (batch_size, 1) :param s: scalar

Gradient clipping.

Predict the next chars given the prefix.

referenc official implementation [insightface](https://github.com/deepinsight/insightface) :param data_set: :param sess: :param embed

Build the VGG 19 Model Parameters ----------- rgb : rgb image placeholder [batch, height, width, 3] values scaled [0, 1]