resnext(cardinality, width, widen_factor = 2, connection = (x, y) -> @. relu(x) + relu(y);
      block_config, nclasses = 1000)

Create a ResNeXt model (reference).

Arguments

• cardinality: the number of groups to use for the convolution
• width: the number of feature maps in each group in the bottleneck
• widen_factor: the factor by which the width of the bottleneck is increased after each stage
• connection: the binary function applied to the output of residual and skip paths in a block
• block_config: a list of the number of residual blocks at each stage
• nclasses: the number of output classes