def test_heatmap_shape():
gcam = GradCAM(paths[1], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert gcam.heatmap.shape == (14, 14)
def test_pred_output_shape():
gcam = GradCAM(paths[0], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert gcam.pred_model(gcam.conv_model(gcam.img)).shape == [1, 1]
def test_heatmap_type():
gcam = GradCAM(paths[1], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert type(gcam.heatmap) == np.ndarray
def test_pred_model_output_type():
gcam = GradCAM(paths[0], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert type(gcam.pred_model(gcam.conv_model(
gcam.img))) == tf.python.framework.ops.EagerTensor
def test_pred_model_type():
gcam = GradCAM(paths[0], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert type(
gcam.conv_model) == tf.python.keras.engine.functional.Functional
def test_image_size():
gcam = GradCAM(paths[0], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert gcam.img.shape[1] == 224
assert gcam.img.shape[2] == 224
def test_image_shape():
gcam = GradCAM(paths[0], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert gcam.img.shape == (1, 224, 224, 3)
def test_img_type():
gcam = GradCAM(paths[0], (224, 224),
model,
last_convolution_layer="conv2d_73")
assert type(gcam.img) == np.ndarray