def test_whenEncoderForward_thenPass(self):
with patch("deepparse.network.nn.LSTM") as lstm_mock:
output_mock = MagicMock()
hidden_mock = MagicMock()
lstm_mock().return_value = output_mock, hidden_mock
with patch("deepparse.network.encoder.pack_padded_sequence"
) as pack_padded_sequence_mock:
packed_sequence_mock = MagicMock()
pack_padded_sequence_mock.return_value = packed_sequence_mock
with patch("deepparse.network.encoder.pad_packed_sequence"
) as pad_packed_sequence_mock:
pad_packed_sequence_mock.return_value = (MagicMock(),
MagicMock())
encoder = Encoder(self.input_size_dim, self.hidden_size,
self.num_layers)
to_predict_mock = MagicMock()
lengths_tensor_mock = MagicMock()
encoder.forward(to_predict_mock, lengths_tensor_mock)
pack_padded_sequence_mock.assert_has_calls([
call(
to_predict_mock,
lengths_tensor_mock.cpu(),
batch_first=True,
enforce_sorted=False,
)
])
lstm_mock.assert_has_calls([call()(packed_sequence_mock)])
class EncoderTest(TestCase):
@classmethod
def setUpClass(cls):
cls.a_torch_device = torch.device("cuda:0")
cls.input_size_dim = 300
cls.hidden_size = 1024
cls.num_layers = 1
def setUp(self) -> None:
self.encoder = Encoder(self.input_size_dim, self.hidden_size,
self.num_layers)
self.encoder.to(self.a_torch_device) # we mount it into the device
self.encoder_input_setUp()
def encoder_input_setUp(self):
# we use the fasttext case since easier
file = open("./tests/network/integration/to_predict_fasttext.p", "rb")
self.to_predict_tensor = pickle.load(file)
self.to_predict_tensor = self.to_predict_tensor.to(self.a_torch_device)
file.close()
self.a_lengths_tensor = torch.tensor([6, 6],
device=self.a_torch_device)
self.a_batch_size = 2
self.max_length = self.a_lengths_tensor[0].item()
def assert_output_is_valid_dim(self, actual_predictions):
for actual_prediction in actual_predictions:
self.assertEqual(self.num_layers, actual_prediction.shape[0])
self.assertEqual(self.a_batch_size, actual_prediction.shape[1])
self.assertEqual(self.hidden_size, actual_prediction.shape[2])
def test_whenForwardStep_thenStepIsOk(self):
predictions = self.encoder.forward(self.to_predict_tensor,
self.a_lengths_tensor)
self.assert_output_is_valid_dim(predictions)