def test_device(self):
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
summary(SingleInputNet(), (1, 28, 28), device=device)
input_data = torch.randn(5, 1, 28, 28)
summary(SingleInputNet(), input_data)
summary(SingleInputNet(), input_data, device=device)
summary(SingleInputNet(), input_data.to(device))
summary(SingleInputNet(), input_data.to(device), device=torch.device("cpu"))
def test_single_input(self, capsys):
model = SingleInputNet()
input_shape = (1, 28, 28)
summary(model, input_shape, depth=1)
verify_output(capsys, "unit_test/test_output/single_input.out")
def test_single_input(capsys: CaptureFixture[str]) -> None:
model = SingleInputNet()
input_shape = (1, 28, 28)
summary(model, input_shape, depth=1)
verify_output(capsys, "unit_test/test_output/single_input.out")
def test_string_result() -> None:
results = summary(SingleInputNet(),
input_size=(16, 1, 28, 28),
verbose=0)
result_str = f"{results}\n"
verify_output_str(result_str, "tests/test_output/single_input.out")
def test_single_input_all_cols(capsys: pytest.CaptureFixture[str]) -> None:
model = SingleInputNet()
col_names = (
"kernel_size",
"input_size",
"output_size",
"num_params",
"mult_adds",
)
input_shape = (7, 1, 28, 28)
summary(model,
input_size=input_shape,
depth=1,
col_names=col_names,
col_width=20)
verify_output(capsys, "tests/test_output/single_input_all.out")
summary(
model,
input_data=torch.randn(*input_shape),
depth=1,
col_names=col_names,
col_width=20,
)
verify_output(capsys, "tests/test_output/single_input_all.out")
def test_single_input(self):
model = SingleInputNet()
input_shape = (1, 28, 28)
results = summary(model, input_shape)
assert results.total_params == 21840
assert results.trainable_params == 21840
def test_string_result(self):
results = summary(SingleInputNet(), (1, 28, 28), verbose=0)
result_str = str(results) + "\n"
with open("unit_test/test_output/single_input.out", encoding="utf-8") as output_file:
expected = output_file.read()
assert result_str == expected
def test_row_settings(capsys: pytest.CaptureFixture[str]) -> None:
model = SingleInputNet()
summary(model,
input_size=(16, 1, 28, 28),
row_settings=("var_names", ))
verify_output(capsys, "tests/test_output/row_settings.out")
def test_single_input(self):
model = SingleInputNet()
input_shape = (1, 28, 28)
results = summary(model, input_shape)
assert len(results.summary_list) == 5, "Should find 6 layers"
assert results.total_params == 21840
assert results.trainable_params == 21840
def test_batch_size_optimization() -> None:
model = SingleInputNet()
# batch size intentionally omitted.
results = summary(model, (1, 28, 28), batch_dim=0)
assert len(results.summary_list) == 5, "Should find 6 layers"
assert results.total_params == 21840
assert results.trainable_params == 21840
def test_single_input() -> None:
model = SingleInputNet()
# input_size keyword arg intentionally omitted.
results = summary(model, (2, 1, 28, 28))
assert len(results.summary_list) == 5, "Should find 6 layers"
assert results.total_params == 21840
assert results.trainable_params == 21840
def test_string_result() -> None:
results = summary(SingleInputNet(),
input_size=(16, 1, 28, 28),
verbose=0)
result_str = str(results) + "\n"
with open("tests/test_output/single_input.out",
encoding="utf-8") as output_file:
expected = output_file.read()
assert result_str == expected
def test_single_input_batch_dim(capsys: CaptureFixture[str]) -> None:
model = SingleInputNet()
input_shape = (7, 1, 28, 28)
summary(model, input_shape, depth=1, batch_dim=None)
verify_output(capsys,
"unit_test/test_output/single_input_batch_dim.out")
input_data = torch.randn(*input_shape)
summary(model, input_data, depth=1, batch_dim=None)
verify_output(capsys,
"unit_test/test_output/single_input_batch_dim.out")
def test_device() -> None:
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = SingleInputNet()
# input_size
summary(model, input_size=(5, 1, 28, 28), device=device)
# input_data
input_data = torch.randn(5, 1, 28, 28)
summary(model, input_data=input_data)
summary(model, input_data=input_data, device=device)
summary(model, input_data=input_data.to(device))
summary(model, input_data=input_data.to(device), device=torch.device("cpu"))
def test_single_input_with_kernel_macs(capsys):
model = SingleInputNet()
input_shape = (1, 28, 28)
summary(
model,
input_shape,
depth=1,
col_names=("kernel_size", "output_size", "num_params", "mult_adds"),
col_width=20,
)
verify_output(capsys, "unit_test/test_output/single_input_all.out")
def test_single_input_batch_dim(
capsys: pytest.CaptureFixture[str]) -> None:
model = SingleInputNet()
col_names = (
"kernel_size",
"input_size",
"output_size",
"num_params",
"mult_adds",
)
summary(
model,
input_size=(1, 28, 28),
depth=1,
col_names=col_names,
col_width=20,
batch_dim=0,
)
verify_output(capsys, "tests/test_output/single_input_batch_dim.out")
def test_basic_summary(capsys: pytest.CaptureFixture[str]) -> None:
model = SingleInputNet()
summary(model)
verify_output(capsys, "unit_test/test_output/basic_summary.out")
def test_single_input(capsys: pytest.CaptureFixture[str]) -> None:
model = SingleInputNet()
summary(model, input_size=(16, 1, 28, 28), depth=1)
verify_output(capsys, "tests/test_output/single_input.out")
def test_input_tensor(self):
input_data = torch.randn(5, 1, 28, 28)
metrics = summary(SingleInputNet(), input_data)
assert metrics.input_size == [torch.Size([5, 1, 28, 28])]
def test_basic_summary(self, capsys):
model = SingleInputNet()
summary(model)
verify_output(capsys, "unit_test/test_output/basic_summary.out")