-
Notifications
You must be signed in to change notification settings - Fork 96
/
test_algebra_onnx_operators.py
206 lines (167 loc) · 7.36 KB
/
test_algebra_onnx_operators.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
import unittest
from distutils.version import StrictVersion
from io import BytesIO
import numpy as np
from numpy.testing import assert_almost_equal
from sklearn.base import BaseEstimator, TransformerMixin
from sklearn.cluster import KMeans
from sklearn.datasets import load_iris
from sklearn.utils.extmath import row_norms
from onnxruntime import InferenceSession
from skl2onnx import convert_sklearn
from skl2onnx.common.data_types import FloatTensorType
from skl2onnx.algebra.onnx_operator import OnnxOperator
from skl2onnx.algebra.onnx_ops import OnnxSub, OnnxDiv
from skl2onnx.algebra.onnx_ops import OnnxReduceSumSquare, OnnxGemm
from skl2onnx.algebra.onnx_ops import OnnxAdd, OnnxArgMin, OnnxSqrt
from onnx import (
helper, TensorProto, load_model,
__version__ as onnx__version__
)
from test_utils import dump_data_and_model
class TestOnnxOperators(unittest.TestCase):
def test_sub(self):
class CustomOpTransformer(BaseEstimator, TransformerMixin):
def __init__(self):
pass
def fit(self, X, y=None):
self.W = np.mean(X, axis=0)
return self
def transform(self, X):
return X - self.W
mat = np.array([[0., 1.], [1., 2.], [3., 4.]])
tr = CustomOpTransformer()
tr.fit(mat)
z = tr.transform(mat)
def conv(scope, operator, container):
W = operator.raw_operator.W
op = OnnxSub(operator.inputs[0], W, output_names=operator.outputs)
op.add_to(scope, container)
text = str(container)
if 'name:"Sub"' not in text:
raise AssertionError(
"Unnamed operator:\n".format(text))
nin = list(op.enumerate_initial_types())
nno = list(op.enumerate_nodes())
nva = list(op.enumerate_variables())
assert len(nin) == 1
assert nin[0][0] == 'input'
assert nin[0][1].shape == [1, 2]
assert len(nno) == 1
assert nno[0].output_names == ['variable']
assert len(nva) == 1
assert isinstance(nva[0], tuple)
assert nva[0][1] == 0
def shape(operator):
N = operator.inputs[0].type.shape[0]
W = operator.raw_operator.W
operator.outputs[0].type.shape = [N, W.shape[0]]
model_onnx = convert_sklearn(
tr, 'a-sub', [('input', FloatTensorType([1, 2]))],
custom_shape_calculators={CustomOpTransformer: shape},
custom_conversion_functions={CustomOpTransformer: conv})
sess = InferenceSession(model_onnx.SerializeToString())
z2 = sess.run(None, {'input': mat.astype(np.float32)})[0]
assert_almost_equal(z, z2)
def test_sub_div(self):
class CustomOpTransformer(BaseEstimator, TransformerMixin):
def __init__(self):
pass
def fit(self, X, y=None):
self.W = np.mean(X, axis=0)
self.S = np.std(X, axis=0)
return self
def transform(self, X):
return (X - self.W) / self.S
mat = np.array([[0., 1.], [0., 1.], [2., 2.]])
tr = CustomOpTransformer()
tr.fit(mat)
z = tr.transform(mat)
def conv(scope, operator, container):
W = operator.raw_operator.W
S = operator.raw_operator.S
X = operator.inputs[0]
out = operator.outputs
op = OnnxDiv(OnnxSub(X, W), S, output_names=out)
op.add_to(scope, container)
def shape(operator):
N = operator.inputs[0].type.shape[0]
W = operator.raw_operator.W
operator.outputs[0].type.shape = [N, W.shape[0]]
model_onnx = convert_sklearn(
tr, 'a-sub-div', [('input', FloatTensorType([1, 2]))],
custom_shape_calculators={CustomOpTransformer: shape},
custom_conversion_functions={CustomOpTransformer: conv})
sess = InferenceSession(model_onnx.SerializeToString())
z2 = sess.run(None, {'input': mat.astype(np.float32)})[0]
assert_almost_equal(z, z2)
def test_sub_kmeans(self):
def conv(scope, operator, container):
X = operator.inputs[0]
out = operator.outputs
op = operator.raw_operator
C = op.cluster_centers_
C2 = row_norms(C, squared=True)
N = X.type.shape[0]
zeros = np.zeros((N, ))
rs = OnnxReduceSumSquare(X, axes=[1], keepdims=1)
z = OnnxAdd(rs, OnnxGemm(X, C, zeros, alpha=-2., transB=1))
y2 = OnnxAdd(C2, z)
lo = OnnxArgMin(y2, axis=1, keepdims=0, output_names=out[:1])
y2s = OnnxSqrt(y2, output_names=out[1:])
lo.add_to(scope, container)
y2s.add_to(scope, container)
data = load_iris()
X = data.data
model = KMeans(n_clusters=3)
model.fit(X)
model_onnx = convert_sklearn(
model, 'a-kmeans', [('input', FloatTensorType([1, X.shape[1]]))],
custom_conversion_functions={KMeans: conv})
dump_data_and_model(X.astype(np.float32)[40:60], model, model_onnx,
basename="SklearnKMeansCustom-Dec4")
def test_unscoped(self):
var2 = OnnxOperator.UnscopedVariable("a")
var1 = OnnxOperator.UnscopedVariable("a")
self.assertEqual(var1, var2)
self.assertEqual(var1, "a")
self.assertEqual(repr(var1), "UnscopedVariable('a')")
def test_constant(self):
cst = OnnxOperator.ConstantVariable("a")
self.assertEqual(cst.value, "a")
@unittest.skipIf(StrictVersion(onnx__version__) < StrictVersion("1.4.0"),
reason="only available for opset >= 10")
def test_onnx_reversed_order(self):
idi = np.identity(2)
idi2 = np.identity(2) * 2
onx = OnnxAdd(OnnxAdd('X', idi), idi2, output_names=['Y'])
model_def = onx.to_onnx({'X': idi.astype(np.float32)})
self.assertEqual(len(model_def.graph.output), 1)
onx = OnnxAdd(idi2, OnnxAdd('X', idi), output_names=['Y'])
model_def = onx.to_onnx({'X': idi.astype(np.float32)})
onnx2 = model_def.SerializeToString()
self.assertEqual(onx.outputs, ['Y'])
# There should be 2 outputs here, bug in ONNX?
self.assertEqual(len(model_def.graph.output), 1)
reload = load_model(BytesIO(onnx2))
self.assertEqual(len(reload.graph.output), 1)
assert reload is not None
def test_onnx_reversed_order_second(self):
X = helper.make_tensor_value_info('X', TensorProto.FLOAT, [2, 2])
Y = helper.make_tensor_value_info('Y', TensorProto.FLOAT, [2, 2])
nodes = [
helper.make_node('Add', ['X', 'idi'], ['temp']),
helper.make_node('Add', ['temp', 'idi2'], ['Y'])
]
graph_def = helper.make_graph(nodes, 't1', [X], [Y])
model_def = helper.make_model(graph_def, producer_name='A')
self.assertEqual(len(model_def.graph.output), 1)
nodes = [
helper.make_node('Add', ['X', 'idi'], ['temp']),
helper.make_node('Add', ['idi2', 'temp'], ['Y'])
]
graph_def = helper.make_graph(nodes, 't1', [X], [Y])
model_def = helper.make_model(graph_def, producer_name='A')
self.assertEqual(len(model_def.graph.output), 1)
if __name__ == "__main__":
unittest.main()