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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
|
import pytest
import numpy
import numpy.testing
from thinc.api import NumpyOps, Model, Linear, noop
from thinc.api import with_array2d, with_array, with_padded, with_list
from thinc.api import with_ragged, with_getitem
from thinc.types import Padded, Ragged
from ..util import get_data_checker
@pytest.fixture(params=[[], [(10, 2)], [(5, 3), (1, 3)], [(2, 3), (0, 3), (1, 3)]])
def shapes(request):
return request.param
@pytest.fixture
def ops():
return NumpyOps()
@pytest.fixture
def list_input(shapes):
data = [numpy.zeros(shape, dtype="f") for shape in shapes]
for i, x in enumerate(data):
# Give values that make it easy to see where rows or columns mismatch.
x += i * 100
x += numpy.arange(x.shape[0]).reshape((-1, 1)) * 10
x += numpy.arange(x.shape[1]).reshape((1, -1))
return data
@pytest.fixture
def ragged_input(ops, list_input):
lengths = numpy.array([len(x) for x in list_input], dtype="i")
if not list_input:
return Ragged(ops.alloc2f(0, 0), lengths)
else:
return Ragged(ops.flatten(list_input), lengths)
@pytest.fixture
def padded_input(ops, list_input):
return ops.list2padded(list_input)
@pytest.fixture
def array_input(ragged_input):
return ragged_input.data
@pytest.fixture
def padded_data_input(padded_input):
x = padded_input
return (x.data, x.size_at_t, x.lengths, x.indices)
@pytest.fixture
def ragged_data_input(ragged_input):
return (ragged_input.data, ragged_input.lengths)
@pytest.fixture
def noop_models():
return [
with_padded(noop()),
with_array(noop()),
with_array2d(noop()),
with_list(noop()),
with_ragged(noop()),
]
# As an example operation, lets just trim the last dimension. That
# should catch stuff that confuses the input and output.
def get_array_model():
def _trim_array_forward(model, X, is_train):
def backprop(dY):
return model.ops.alloc2f(dY.shape[0], dY.shape[1] + 1)
return X[:, :-1], backprop
return with_array2d(Model("trimarray", _trim_array_forward))
def get_list_model():
def _trim_list_forward(model, Xs, is_train):
def backprop(dYs):
dXs = []
for dY in dYs:
dXs.append(model.ops.alloc2f(dY.shape[0], dY.shape[1] + 1))
return dXs
Ys = [X[:, :-1] for X in Xs]
return Ys, backprop
return with_list(Model("trimlist", _trim_list_forward))
def get_padded_model():
def _trim_padded_forward(model, Xp, is_train):
def backprop(dYp):
dY = dYp.data
dX = model.ops.alloc3f(dY.shape[0], dY.shape[1], dY.shape[2] + 1)
return Padded(dX, dYp.size_at_t, dYp.lengths, dYp.indices)
assert isinstance(Xp, Padded)
X = Xp.data
X = X.reshape((X.shape[0] * X.shape[1], X.shape[2]))
X = X[:, :-1]
X = X.reshape((Xp.data.shape[0], Xp.data.shape[1], X.shape[1]))
return Padded(X, Xp.size_at_t, Xp.lengths, Xp.indices), backprop
return with_padded(Model("trimpadded", _trim_padded_forward))
def get_ragged_model():
def _trim_ragged_forward(model, Xr, is_train):
def backprop(dYr):
dY = dYr.data
dX = model.ops.alloc2f(dY.shape[0], dY.shape[1] + 1)
return Ragged(dX, dYr.lengths)
return Ragged(Xr.data[:, :-1], Xr.lengths), backprop
return with_ragged(Model("trimragged", _trim_ragged_forward))
def check_initialize(model, inputs):
# Just check that these run and don't hit errors. I guess we should add a
# spy and check that model.layers[0].initialize gets called, but shrug?
model.initialize()
model.initialize(X=inputs)
model.initialize(X=inputs, Y=model.predict(inputs))
def check_transform_produces_correct_output_type_forward(model, inputs, checker):
# It's pretty redundant to check these three assertions, so if the tests
# get slow this could be removed. I think it should be fine though?
outputs = model.predict(inputs)
assert checker(inputs, outputs)
outputs, _ = model(inputs, is_train=True)
assert checker(inputs, outputs)
outputs, _ = model(inputs, is_train=False)
assert checker(inputs, outputs)
def check_transform_produces_correct_output_type_backward(model, inputs, checker):
# It's pretty redundant to check these three assertions, so if the tests
# get slow this could be removed. I think it should be fine though?
outputs, backprop = model.begin_update(inputs)
d_inputs = backprop(outputs)
assert checker(inputs, d_inputs)
def check_transform_doesnt_change_noop_values(model, inputs, d_outputs):
# Check that if we're wrapping a noop() layer in the transform, we don't
# change the output values.
outputs, backprop = model.begin_update(inputs)
d_inputs = backprop(d_outputs)
if isinstance(outputs, list):
for i in range(len(outputs)):
numpy.testing.assert_equal(inputs[i], outputs[i])
numpy.testing.assert_equal(d_outputs[i], d_inputs[i])
elif isinstance(outputs, numpy.ndarray):
numpy.testing.assert_equal(inputs, outputs)
numpy.testing.assert_equal(d_outputs, d_inputs)
elif isinstance(outputs, Ragged):
numpy.testing.assert_equal(inputs.data, outputs.data)
numpy.testing.assert_equal(d_outputs.data, d_inputs.data)
elif isinstance(outputs, Padded):
numpy.testing.assert_equal(inputs.data, outputs.data)
numpy.testing.assert_equal(d_inputs.data, d_inputs.data)
def test_noop_transforms(noop_models, ragged_input, padded_input, list_input):
# Make distinct backprop values,
# to check that the gradients get passed correctly
d_ragged = Ragged(ragged_input.data + 1, ragged_input.lengths)
d_padded = padded_input.copy()
d_padded.data += 1
d_list = [dx + 1 for dx in list_input]
for model in noop_models:
print(model.name)
check_transform_doesnt_change_noop_values(model, padded_input, d_padded)
check_transform_doesnt_change_noop_values(model, list_input, d_list)
check_transform_doesnt_change_noop_values(model, ragged_input, d_ragged)
def test_with_array_initialize(ragged_input, padded_input, list_input, array_input):
for inputs in (ragged_input, padded_input, list_input, array_input):
check_initialize(get_array_model(), inputs)
def test_with_padded_initialize(
ragged_input, padded_input, list_input, padded_data_input
):
for inputs in (ragged_input, padded_input, list_input, padded_data_input):
check_initialize(get_padded_model(), inputs)
def test_with_list_initialize(ragged_input, padded_input, list_input):
for inputs in (ragged_input, padded_input, list_input):
check_initialize(get_list_model(), inputs)
def test_with_ragged_initialize(
ragged_input, padded_input, list_input, ragged_data_input
):
for inputs in (ragged_input, padded_input, list_input, ragged_data_input):
check_initialize(get_ragged_model(), inputs)
def test_with_array_forward(ragged_input, padded_input, list_input, array_input):
for inputs in (ragged_input, padded_input, list_input, array_input):
checker = get_data_checker(inputs)
model = get_array_model()
check_transform_produces_correct_output_type_forward(model, inputs, checker)
def test_with_list_forward(ragged_input, padded_input, list_input):
for inputs in (ragged_input, padded_input, list_input):
checker = get_data_checker(inputs)
model = get_list_model()
check_transform_produces_correct_output_type_forward(model, inputs, checker)
def test_with_padded_forward(ragged_input, padded_input, list_input, padded_data_input):
for inputs in (ragged_input, padded_input, list_input, padded_data_input):
checker = get_data_checker(inputs)
model = get_padded_model()
check_transform_produces_correct_output_type_forward(model, inputs, checker)
def test_with_ragged_forward(ragged_input, padded_input, list_input, ragged_data_input):
for inputs in (ragged_input, padded_input, list_input, ragged_data_input):
checker = get_data_checker(inputs)
model = get_ragged_model()
check_transform_produces_correct_output_type_forward(model, inputs, checker)
def test_with_array_backward(ragged_input, padded_input, list_input, array_input):
for inputs in (ragged_input, padded_input, list_input, array_input):
checker = get_data_checker(inputs)
model = get_array_model()
check_transform_produces_correct_output_type_backward(model, inputs, checker)
def test_with_list_backward(ragged_input, padded_input, list_input):
for inputs in (ragged_input, padded_input, list_input):
checker = get_data_checker(inputs)
model = get_list_model()
check_transform_produces_correct_output_type_backward(model, inputs, checker)
def test_with_ragged_backward(
ragged_input, padded_input, list_input, ragged_data_input
):
for inputs in (ragged_input, padded_input, list_input, ragged_data_input):
checker = get_data_checker(inputs)
model = get_ragged_model()
check_transform_produces_correct_output_type_backward(model, inputs, checker)
def test_with_padded_backward(
ragged_input, padded_input, list_input, padded_data_input
):
for inputs in (ragged_input, padded_input, list_input, padded_data_input):
checker = get_data_checker(inputs)
model = get_padded_model()
check_transform_produces_correct_output_type_backward(model, inputs, checker)
def test_with_getitem():
data = (
numpy.asarray([[1, 2, 3, 4]], dtype="f"),
numpy.asarray([[5, 6, 7, 8]], dtype="f"),
)
model = with_getitem(1, Linear())
model.initialize(data, data)
Y, backprop = model.begin_update(data)
assert len(Y) == len(data)
assert numpy.array_equal(Y[0], data[0]) # the other item stayed the same
assert not numpy.array_equal(Y[1], data[1])
dX = backprop(Y)
assert numpy.array_equal(dX[0], data[0])
assert not numpy.array_equal(dX[1], data[1])
|
