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
|
"""Tests for running inplace prediction."""
from concurrent.futures import ThreadPoolExecutor
import numpy as np
import pandas as pd
import pytest
from scipy import sparse
import xgboost as xgb
from xgboost import testing as tm
from xgboost.testing.data import np_dtypes, pd_dtypes
from xgboost.testing.shared import validate_leaf_output
def run_threaded_predict(X, rows, predict_func):
results = []
per_thread = 20
with ThreadPoolExecutor(max_workers=10) as e:
for i in range(0, rows, int(rows / per_thread)):
if hasattr(X, "iloc"):
predictor = X.iloc[i : i + per_thread, :]
else:
predictor = X[i : i + per_thread, ...]
f = e.submit(predict_func, predictor)
results.append(f)
for f in results:
assert f.result()
def run_predict_leaf(device: str) -> np.ndarray:
rows = 100
cols = 4
classes = 5
num_parallel_tree = 4
num_boost_round = 10
rng = np.random.RandomState(1994)
X = rng.randn(rows, cols)
y = rng.randint(low=0, high=classes, size=rows)
m = xgb.DMatrix(X, y)
booster = xgb.train(
{
"num_parallel_tree": num_parallel_tree,
"num_class": classes,
"tree_method": "hist",
},
m,
num_boost_round=num_boost_round,
)
booster.set_param({"device": device})
empty = xgb.DMatrix(np.ones(shape=(0, cols)))
empty_leaf = booster.predict(empty, pred_leaf=True)
assert empty_leaf.shape[0] == 0
leaf = booster.predict(m, pred_leaf=True, strict_shape=True)
assert leaf.shape[0] == rows
assert leaf.shape[1] == num_boost_round
assert leaf.shape[2] == classes
assert leaf.shape[3] == num_parallel_tree
validate_leaf_output(leaf, num_parallel_tree)
n_iters = np.int32(2)
sliced = booster.predict(
m,
pred_leaf=True,
iteration_range=(0, n_iters),
strict_shape=True,
)
first = sliced[0, ...]
assert np.prod(first.shape) == classes * num_parallel_tree * n_iters
# When there's only 1 tree, the output is a 1 dim vector
booster = xgb.train({"tree_method": "hist"}, num_boost_round=1, dtrain=m)
booster.set_param({"device": device})
assert booster.predict(m, pred_leaf=True).shape == (rows,)
return leaf
def test_predict_leaf() -> None:
run_predict_leaf("cpu")
def test_predict_shape():
from sklearn.datasets import fetch_california_housing
X, y = fetch_california_housing(return_X_y=True)
reg = xgb.XGBRegressor(n_estimators=1)
reg.fit(X, y)
predt = reg.get_booster().predict(xgb.DMatrix(X), strict_shape=True)
assert len(predt.shape) == 2
assert predt.shape[0] == X.shape[0]
assert predt.shape[1] == 1
contrib = reg.get_booster().predict(
xgb.DMatrix(X), pred_contribs=True, strict_shape=True
)
assert len(contrib.shape) == 3
assert contrib.shape[1] == 1
contrib = reg.get_booster().predict(
xgb.DMatrix(X), pred_contribs=True, approx_contribs=True
)
assert len(contrib.shape) == 2
assert contrib.shape[1] == X.shape[1] + 1
interaction = reg.get_booster().predict(
xgb.DMatrix(X), pred_interactions=True, approx_contribs=True
)
assert len(interaction.shape) == 3
assert interaction.shape[1] == X.shape[1] + 1
assert interaction.shape[2] == X.shape[1] + 1
interaction = reg.get_booster().predict(
xgb.DMatrix(X), pred_interactions=True, approx_contribs=True, strict_shape=True
)
assert len(interaction.shape) == 4
assert interaction.shape[1] == 1
assert interaction.shape[2] == X.shape[1] + 1
assert interaction.shape[3] == X.shape[1] + 1
class TestInplacePredict:
"""Tests for running inplace prediction"""
@classmethod
def setup_class(cls):
cls.rows = 1000
cls.cols = 10
cls.missing = 11 # set to integer for testing
cls.rng = np.random.RandomState(1994)
cls.X = cls.rng.randn(cls.rows, cls.cols)
missing_idx = [i for i in range(0, cls.cols, 4)]
cls.X[:, missing_idx] = cls.missing # set to be missing
cls.y = cls.rng.randn(cls.rows)
dtrain = xgb.DMatrix(cls.X, cls.y)
cls.test = xgb.DMatrix(cls.X[:10, ...], missing=cls.missing)
cls.num_boost_round = 10
cls.booster = xgb.train({"tree_method": "hist"}, dtrain, num_boost_round=10)
def test_predict(self):
booster = self.booster
X = self.X
test = self.test
predt_from_array = booster.inplace_predict(X[:10, ...], missing=self.missing)
predt_from_dmatrix = booster.predict(test)
X_obj = X.copy().astype(object)
assert X_obj.dtype.hasobject is True
assert X.dtype.hasobject is False
np.testing.assert_allclose(
booster.inplace_predict(X_obj), booster.inplace_predict(X)
)
np.testing.assert_allclose(predt_from_dmatrix, predt_from_array)
predt_from_array = booster.inplace_predict(
X[:10, ...], iteration_range=(0, 4), missing=self.missing
)
predt_from_dmatrix = booster.predict(test, iteration_range=(0, 4))
np.testing.assert_allclose(predt_from_dmatrix, predt_from_array)
with pytest.raises(ValueError):
booster.predict(test, iteration_range=(0, booster.num_boosted_rounds() + 2))
default = booster.predict(test)
range_full = booster.predict(test, iteration_range=(0, self.num_boost_round))
np.testing.assert_allclose(range_full, default)
range_full = booster.predict(
test, iteration_range=(0, booster.num_boosted_rounds())
)
np.testing.assert_allclose(range_full, default)
def predict_dense(x):
inplace_predt = booster.inplace_predict(x)
d = xgb.DMatrix(x)
copied_predt = booster.predict(d)
return np.all(copied_predt == inplace_predt)
for i in range(10):
run_threaded_predict(X, self.rows, predict_dense)
def predict_csr(x):
inplace_predt = booster.inplace_predict(sparse.csr_matrix(x))
d = xgb.DMatrix(x)
copied_predt = booster.predict(d)
return np.all(copied_predt == inplace_predt)
for i in range(10):
run_threaded_predict(X, self.rows, predict_csr)
@pytest.mark.skipif(**tm.no_pandas())
def test_predict_pd(self):
X = self.X
# construct it in column major style
df = pd.DataFrame({str(i): X[:, i] for i in range(X.shape[1])})
booster = self.booster
df_predt = booster.inplace_predict(df)
arr_predt = booster.inplace_predict(X)
dmat_predt = booster.predict(xgb.DMatrix(X))
X = df.values
X = np.asfortranarray(X)
fort_predt = booster.inplace_predict(X)
np.testing.assert_allclose(dmat_predt, arr_predt)
np.testing.assert_allclose(df_predt, arr_predt)
np.testing.assert_allclose(fort_predt, arr_predt)
def test_base_margin(self):
booster = self.booster
base_margin = self.rng.randn(self.rows)
from_inplace = booster.inplace_predict(data=self.X, base_margin=base_margin)
dtrain = xgb.DMatrix(self.X, self.y, base_margin=base_margin)
from_dmatrix = booster.predict(dtrain)
np.testing.assert_allclose(from_dmatrix, from_inplace)
@pytest.mark.skipif(**tm.no_pandas())
def test_dtypes(self) -> None:
for orig, x in np_dtypes(self.rows, self.cols):
predt_orig = self.booster.inplace_predict(orig)
predt = self.booster.inplace_predict(x)
np.testing.assert_allclose(predt, predt_orig)
# unsupported types
for dtype in [
np.bytes_,
np.complex64,
np.complex128,
]:
X: np.ndarray = np.array(orig, dtype=dtype)
with pytest.raises(ValueError):
self.booster.inplace_predict(X)
@pytest.mark.skipif(**tm.no_pandas())
def test_pd_dtypes(self) -> None:
from pandas.api.types import is_bool_dtype
for orig, x in pd_dtypes():
dtypes = orig.dtypes if isinstance(orig, pd.DataFrame) else [orig.dtypes]
if isinstance(orig, pd.DataFrame) and is_bool_dtype(dtypes[0]):
continue
y = np.arange(x.shape[0])
Xy = xgb.DMatrix(orig, y, enable_categorical=True)
booster = xgb.train({"tree_method": "hist"}, Xy, num_boost_round=1)
predt_orig = booster.inplace_predict(orig)
predt = booster.inplace_predict(x)
np.testing.assert_allclose(predt, predt_orig)
|
