def test_check_input6():
with pytest.raises(ValueError,
match='Please make sure that the directory' +
' where you want to save the models is empty!'):
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
validation_metrics, '.', params)
trainer.train()
def test_check_input2():
with pytest.raises(
TypeError,
match='Please check you are using the right model object,' +
' or the right order of the attributes!'):
trainer = SingleObjectiveTrainer(dataHandler, None, correctness_loss,
validation_metrics, save_to_path,
params)
trainer.train()
def test_check_input9():
with pytest.raises(
TypeError,
match=
'Please make sure that the optimizer is a pytorch Optimizer object!'
):
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
validation_metrics, save_to_path,
params, model)
trainer.train()
def test_check_input5():
with pytest.raises(
TypeError,
match='Please check you are using the right metric objects,' +
' or the right order of the attributes!'):
validation_metrics_tmp = validation_metrics.copy()
validation_metrics_tmp[0] = model
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
validation_metrics_tmp, save_to_path,
params)
trainer.train()
def test_check_input3():
class TestModel(nn.Module):
def forward(self):
return 1
with pytest.raises(
TypeError,
match=
'Please check if your models has initialize_model\\(\\) method defined!'
):
trainer = SingleObjectiveTrainer(dataHandler, TestModel(),
correctness_loss, validation_metrics,
save_to_path, params)
trainer.train()
def test_init_objects():
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
validation_metrics, save_to_path, params)
assert type(trainer._train_dataloader) == DataLoader
assert type(trainer.pareto_manager) == ParetoManager
assert trainer.pareto_manager.path == save_to_path
assert type(trainer.validator) == Validator
def test_read_yaml_params():
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
validation_metrics, save_to_path, params)
assert trainer.seed == 42
assert trainer.learning_rate == 1e-3
assert trainer.batch_size_training == 500
assert trainer.shuffle_training is True
assert trainer.drop_last_batch_training is True
assert trainer.batch_size_validation == 500
assert trainer.shuffle_validation is True
assert trainer.drop_last_batch_validation is False
assert trainer.number_of_epochs == 50
assert trainer.anneal is True
assert trainer.beta_start == 0
assert trainer.beta_cap == 0.3
assert trainer.beta_step == 0.3 / 10000
def test_check_input8():
# check for None metrics
with pytest.raises(ValueError,
match='The validation_metrics are None,' +
' please make sure to give valid validation_metrics!'):
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
None, save_to_path, params)
trainer.train()
# check if length is at least 1
validation_metrics_tmp = []
with pytest.raises(
ValueError,
match=
'Please check you have defined at least one validation metric!'):
trainer = SingleObjectiveTrainer(dataHandler, model, correctness_loss,
validation_metrics_tmp, save_to_path,
params)
trainer.train()
test_input_data_path = os.path.join(dir_path, 'movielens_small_test_input.npy')
test_output_data_path = os.path.join(dir_path, 'movielens_small_test_test.npy')
products_data_path = os.path.join(dir_path, 'movielens_products_data.npy')
data_handler = AEDataHandler('MovieLensSmall', train_data_path,
validation_input_data_path,
validation_output_data_path, test_input_data_path,
test_output_data_path)
input_dim = data_handler.get_input_dim()
output_dim = data_handler.get_output_dim()
products_data_np = np.load(products_data_path)
products_data_torch = torch.tensor(products_data_np,
dtype=torch.float32).to(device)
# create model
model = MultiVAE(params='yaml_files/params_multi_VAE_training.yaml')
# correctnes loss
loss = VAELoss()
recallAtK = RecallAtK(k=10)
revenueAtK = RevenueAtK(k=10, revenue=products_data_np)
validation_metrics = [recallAtK, revenueAtK]
trainer = SingleObjectiveTrainer(data_handler, model, loss, validation_metrics,
save_to_path)
trainer.train()
print(trainer.pareto_manager._pareto_front)