ResNeXt

resnext.py - academic (idiomatic)
resnext_cifar10.py - academic (idiomatic)
resnext_c.py - production (composable)

Paper

Macro-Architecture

Micro-Architecture

Stem Group

ResNeXt Block with Identity Shortcut

ResNeXt Block with Projection Shortcut

Cardinality

Classifier

Composable

Example: Instantiate a stock ResNeXt model

from resnext_c import ResNeXt
# ResNeXt50 from research paper
resnext = ResNeXt(50)

# ResNeXt50 custom input shape/classes
resnext = ResNeXt(50, input_shape=(128, 128, 3), n_classes=50)

# getter for the tf.keras model
model = resnext.model

Example: Compose and Train a ResNext model

 ''' Example for constructing/training a ResNeXt model on CIFAR-10
    '''
    # Example of constructing a mini-ResNeXt
    groups =  [ { 'filters_in': 128,  'filters_out' : 256,  'n_blocks': 1 },
                { 'filters_in': 256,  'filters_out' : 512,  'n_blocks': 2 },
                { 'filters_in': 512,  'filters_out' : 1024, 'n_blocks': 2 } ]
    resnext = ResNeXt(groups, input_shape=(32, 32, 3), n_classes=10)
    resnext.model.summary()
    resnext.cifar10()