Stack n bottleneck modules where n is inferred from the depth of the network. Args: block: block type used to construct ResNext

Computes the precision@k for the specified values of k

Stack n bottleneck modules where n is inferred from the depth of the network. Args: name: string name of the current block.

3x3 convolution with padding

3x3 convolution with padding

make dir if not exist

Constructor Args: baseWidth: baseWidth for ResNeXt. cardinality: number of convolution groups. layers: co

Constructor Args: cardinality: number of convolution groups. depth: number of layers. num_classes: number

Loads checkpoints from inputs and returns a model with averaged weights. Args: inputs: An iterable of string paths of checkpoints to load f

Initialize parameters following the way proposed in the paper.

paths is a distionary with {name:filepath} pair

Constructor Args: inplanes: input channel dimensionality planes: output channel dimensionality baseWidth:

Constructor Args: in_channels: input channel dimensionality out_channels: output channel dimensionality s