gobbli.dataset.base module

class gobbli.dataset.base.BaseDataset(*args, **kwargs)[source]

Bases: abc.ABC

Abstract base class for datasets used for benchmarking and testing.

Derived classes should account for the following:
  • Dataset order should be consistent so limiting can work correctly

Blank constructor needed to satisfy mypy

abstract X_test()[source]
abstract X_train()[source]
classmethod data_dir()[source]
Return type

Path

embed_input(embed_batch_size=32, pooling=<EmbedPooling.MEAN: 'mean'>, limit=None)[source]
Return type

EmbedInput

classmethod load(*args, **kwargs)[source]
Return type

BaseDataset

predict_input(predict_batch_size=32, limit=None)[source]
Return type

PredictInput

train_input(train_batch_size=32, valid_batch_size=8, num_train_epochs=3, valid_proportion=0.2, split_seed=1234, shuffle_seed=1234, limit=None)[source]
Return type

TrainInput

abstract y_test()[source]
abstract y_train()[source]
gobbli.dataset.base.LOGGER = <Logger gobbli.dataset.base (WARNING)>
gobbli.dataset.base.dataset_dir()[source]
Return type

Path