def test_shape(self):
paddle_model = mobilenet_v2(pretrained=True)
img_path = 'imgs/catdog.jpg'
algo = it.GradShapCVInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, resize_to=256, crop_to=224, visual=False)
result = np.array([*exp.shape])
assert_arrays_almost_equal(self, result, np.array([1, 3, 224, 224]))
def test_cv_multiple_inputs(self):
np.random.seed(42)
paddle_model = mobilenet_v2(pretrained=True)
img_path = ['imgs/catdog.jpg', 'imgs/catdog.jpg']
algo = it.GradShapCVInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, resize_to=256, crop_to=224, visual=False, n_samples=1)
result = np.array([*exp.shape])
assert_arrays_almost_equal(self, result, np.array([2, 3, 224, 224]))
def test_algo(self):
np.random.seed(42)
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0, 255, size=(1, 64, 64, 3), dtype=np.uint8)
algo = it.GradShapCVInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, visual=False)
result = np.array([exp.mean(), exp.std(), exp.min(), exp.max()])
desired = np.array([ 3.4148130e-07, 2.6306865e-05, -2.3330076e-04, 3.4065323e-04])
assert_arrays_almost_equal(self, result, desired)
def test_case2(self):
paddle.disable_static()
model = mobilenet_v2()
predictor = TableLatencyPredictor(table_file='SD710')
model_file, param_file = save_cls_model(model,
input_shape=[1, 3, 224, 224],
save_dir="./inference_model",
data_type='fp32')
latency = predictor.predict(model_file=model_file,
param_file=param_file,
data_type='fp32')
assert latency > 0
def test_algo_layer(self):
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0,
255,
size=(1, 64, 64, 3),
dtype=np.uint8)
algo = it.GradCAMInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, 'features.16.conv.3', visual=False)
result = np.array([exp.mean(), exp.std(), exp.min(), exp.max()])
desired = np.array(
[1.07594451e-03, 4.04298247e-04, 5.23229188e-04, 1.60590897e-03])
assert_arrays_almost_equal(self, result, desired)
def test_cv_multiple_inputs(self):
paddle_model = mobilenet_v2(pretrained=True)
img_path = ['imgs/catdog.jpg', 'imgs/catdog.jpg']
algo = it.GradCAMInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path,
'features.18.2',
resize_to=256,
crop_to=224,
visual=False)
result = np.array([*exp.shape])
desired = np.array([2, 7, 7])
assert_arrays_almost_equal(self, result, desired)
def test_algo(self):
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0,
255,
size=(1, 64, 64, 3),
dtype=np.uint8)
algo = it.SmoothGradInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, n_samples=5, visual=False)
result = np.array([exp.mean(), exp.std(), exp.min(), exp.max()])
desired = np.array(
[5.1930110e-06, 4.0496187e-03, -2.3638349e-02, 3.2775488e-02])
assert_arrays_almost_equal(self, result, desired)
def test_save(self):
import matplotlib
matplotlib.use('agg') # non-GUI, for skipping.
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0, 255, size=(1, 64, 64, 3), dtype=np.uint8)
algo = it.GradShapCVInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, visual=True, save_path='tmp.jpg')
result = np.array([exp.mean(), exp.std(), exp.min(), exp.max()])
desired = np.array([ 3.4148130e-07, 2.6306865e-05, -2.3330076e-04, 3.4065323e-04])
assert_arrays_almost_equal(self, result, desired)
os.remove('tmp.jpg')
def test_shape(self):
paddle_model = mobilenet_v2(pretrained=True)
img_path = 'imgs/catdog.jpg'
algo = it.ScoreCAMInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path,
'features.18.2',
resize_to=64,
crop_to=64,
visual=False)
result = np.array([*exp.shape])
desired = np.array([1., 64., 64.])
assert_arrays_almost_equal(self, result, desired)
def test_algo(self):
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0,
255,
size=(1, 64, 64, 3),
dtype=np.uint8)
algo = it.GradCAMInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path, 'features.18.2', visual=False)
result = np.array([exp.mean(), exp.std(), exp.min(), exp.max()])
desired = np.array(
[2.4055550e-04, 9.1252205e-06, 2.2793890e-04, 2.5061620e-04],
dtype=np.float32)
assert_arrays_almost_equal(self, result, desired)
def test_visual(self):
import matplotlib
matplotlib.use('agg') # non-GUI, for skipping.
paddle_model = mobilenet_v2(pretrained=True)
img_path = 'imgs/catdog.jpg'
algo = it.SmoothGradInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path,
n_samples=1,
resize_to=256,
crop_to=224,
visual=True,
save_path='tmp.jpg')
result = np.array([*exp.shape])
assert_arrays_almost_equal(self, result, np.array([1, 3, 224, 224]))
os.remove('tmp.jpg')
def test_case3(self):
paddle.disable_static()
model = mobilenet_v2()
model_file, param_file = save_cls_model(model,
input_shape=[1, 3, 224, 224],
save_dir="./inference_model",
data_type='fp32')
pbmodel_file = opt_model(model_file=model_file,
param_file=param_file,
optimize_out_type='protobuf')
pred = LatencyPredictor()
paddle.enable_static()
with open(pbmodel_file, "rb") as f:
fluid_program = paddle.fluid.framework.Program.parse_from_string(
f.read())
graph = paddleslim.core.GraphWrapper(fluid_program)
graph_keys = pred._get_key_info_from_graph(graph=graph)
assert len(graph_keys) > 0
def test_algo(self):
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0,
255,
size=(1, 64, 64, 3),
dtype=np.uint8)
algo = it.ScoreCAMInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path,
'features.16.conv.3',
resize_to=64,
crop_to=64,
visual=False)
result = np.array([exp.mean(), exp.std(), exp.min(), exp.max()])
desired = np.array([2.3779001, 1.80234987, -0.33591489, 5.09665543])
assert_arrays_almost_equal(self, result, desired)
def test_save(self):
import matplotlib
matplotlib.use('agg') # non-GUI, for skipping.
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0,
255,
size=(1, 64, 64, 3),
dtype=np.uint8)
algo = it.LIMECVInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path,
num_samples=42,
batch_size=20,
resize_to=64,
crop_to=64,
visual=True,
save_path='tmp.jpg')
os.remove('tmp.jpg')
def test_case10(self):
paddle.disable_static()
model = mobilenet_v2()
model2 = ModelCase6()
model3 = ModelCase7()
predictor = TableLatencyPredictor(table_file='SD710')
model_file, param_file = save_cls_model(model,
input_shape=[1, 3, 250, 250],
save_dir="./inference_model",
data_type='fp32')
latency = predictor.predict(model_file=model_file,
param_file=param_file,
data_type='fp32')
assert latency > 0
model_file, param_file = save_cls_model(model,
input_shape=[1, 3, 250, 250],
save_dir="./inference_model",
data_type='int8')
latency = predictor.predict(model_file=model_file,
param_file=param_file,
data_type='int8')
assert latency > 0
model_file, param_file = save_cls_model(model2,
input_shape=[1, 3, 16, 16],
save_dir="./inference_model",
data_type='fp32')
latency = predictor.predict(model_file=model_file,
param_file=param_file,
data_type='fp32')
assert latency > 0
model_file, param_file = save_det_model(model3,
input_shape=[1, 255, 14, 14],
save_dir="./inference_model",
data_type='fp32')
latency = predictor.predict(model_file=model_file,
param_file=param_file,
data_type='fp32')
assert latency > 0
def test_shape_and_algo(self):
paddle_model = mobilenet_v2(pretrained=True)
np.random.seed(42)
img_path = np.random.randint(0,
255,
size=(1, 64, 64, 3),
dtype=np.uint8)
algo = it.LIMECVInterpreter(paddle_model, device='cpu')
exp = algo.interpret(img_path,
num_samples=100,
batch_size=50,
resize_to=64,
crop_to=64,
visual=False)
k = list(exp.keys())[0]
result = np.zeros(len(exp[k]))
for sp_id, v in exp[k]:
result[sp_id] = v
result = np.array([k, *result])
desired = np.array([
4.90000000e+02, 1.86663490e-03, -6.07835046e-02, 3.15221286e-02,
-4.33815891e-02, 5.98708568e-02, -4.03566565e-02, 2.32181935e-04,
-1.88198336e-02, -5.48955918e-04, -3.05450687e-02, 6.40058578e-02,
6.46084885e-02, -1.35210916e-03, -3.33294607e-02, -3.67286859e-02,
3.36535163e-02, -8.59251274e-03, 4.58715622e-02, 4.07490782e-02,
1.16548644e-01, 8.28152785e-02, 6.30378504e-02, -9.37234923e-03,
5.44956613e-03, 2.78750946e-03, -5.23468125e-02, 2.56323242e-03,
9.42740024e-02, 7.96398789e-02, 6.01124075e-02, -5.76994244e-02,
7.03640418e-02, 3.61672162e-02, 4.97282116e-02, 5.11243389e-02,
1.89338674e-01, 1.21587496e-02, 4.50969712e-02, 4.63132584e-02,
-4.05242203e-03, 5.21016618e-02, 7.57953552e-02, 2.91536810e-02,
1.51803549e-02, 6.08129130e-02, 5.33129929e-02, -2.09711908e-02,
-4.93060550e-03, -1.41900513e-02, 7.03506893e-02, 5.65046961e-02,
-2.94891549e-02, -3.20462166e-02, 2.93698146e-02, -1.15419265e-01,
-8.70454571e-03, -3.36799397e-02, 2.18574095e-02, 1.36021779e-02,
4.22237924e-02, -7.91892625e-02, 1.59767297e-02, 2.24134222e-02
])
assert_arrays_almost_equal(self, result, desired)
