def test_basic_functionality_with_FITBClosedVocabGGNN(self):
make_tasks_and_preprocess(seed=514,
dataset_name=self.dataset_name,
experiment_name=self.experiment_name,
n_jobs=30,
task_names=['FITBTask'],
model_names_labels_and_prepro_kwargs=[
('FITBClosedVocabGGNN', 'test', frozenset(), dict(), dict(max_nodes_per_graph=50)),
(
'FITBClosedVocabGGNN', 'test2', frozenset(), dict(),
dict(max_nodes_per_graph=50)),
],
test=True)
make_tasks_and_preprocess(seed=514,
dataset_name=self.dataset_name,
experiment_name=self.experiment_name,
n_jobs=30,
task_names=['FITBTask'],
model_names_labels_and_prepro_kwargs=[
(
'FITBClosedVocabGGNN', 'test3', frozenset(), dict(),
dict(max_nodes_per_graph=50)),
],
skip_make_tasks=True,
test=True)
def test_basic_functionality_with_FITBClosedVocabGGNN(self):
make_tasks_and_preprocess(seed=514,
dataset_name=self.dataset_name,
experiment_name=self.experiment_name,
n_jobs=30,
task_names=['FITBTask'],
model_names_labels_and_prepro_kwargs=[
('FITBClosedVocabGGNN', 'all_edge', frozenset(), dict(),
dict(max_nodes_per_graph=50))],
test=True)
train_model_for_experiment(dataset_name=self.dataset_name,
experiment_name=self.experiment_name,
experiment_run_log_id='test_log_id',
seed=5145,
gpu_ids=(0, 1),
model_name='FITBClosedVocabGGNN',
model_label='all_edge',
model_kwargs=dict(hidden_size=17,
type_emb_size=7,
name_emb_size=5,
n_msg_pass_iters=2),
init_fxn_name='Xavier',
init_fxn_kwargs=dict(),
loss_fxn_name='FITBLoss',
loss_fxn_kwargs=dict(),
optimizer_name='Adam',
optimizer_kwargs={'learning_rate': .0002},
val_fraction=0.15,
n_workers=4,
n_epochs=2,
evaluation_metrics=('evaluate_FITB_accuracy',),
n_batch=64,
test=True)
