cogdl.trainers.deepergcn_trainer

Module Contents

Classes

DeeperGCNTrainer

Helper class that provides a standard way to create an ABC using

Functions

random_partition_graph(num_nodes, cluster_number=10)
generate_subgraphs(edge_index, parts, cluster_number=10, batch_size=1)

cogdl.trainers.deepergcn_trainer.random_partition_graph(num_nodes, cluster_number=10)

cogdl.trainers.deepergcn_trainer.generate_subgraphs(edge_index, parts, cluster_number=10, batch_size=1)

class cogdl.trainers.deepergcn_trainer.DeeperGCNTrainer(args)

Bases: cogdl.trainers.base_trainer.BaseTrainer

Helper class that provides a standard way to create an ABC using inheritance.

fit(self, model, data)

test_gpu_volume(self)

_train_step(self)

_test_step(self, split='val')

loss(self, data)

predict(self, data)

classmethod build_trainer_from_args(cls, args)

Build a new trainer instance.