Skip to content

Commit f729d69

Browse files
miguelusquemiguelusque
committed
Fix 270. Move imports to the top.
1 parent 1e37447 commit f729d69

9 files changed

+303
-284
lines changed

‎cuml/coordinate_descent_demo.ipynb‎

Lines changed: 59 additions & 59 deletions
Original file line numberDiff line numberDiff line change
@@ -28,7 +28,7 @@
2828
"cell_type": "markdown",
2929
"metadata": {},
3030
"source": [
31-
"## Define Parameters"
31+
"## Imports"
3232
]
3333
},
3434
{
@@ -37,19 +37,19 @@
3737
"metadata": {},
3838
"outputs": [],
3939
"source": [
40-
"n_samples = 2**17\n",
41-
"n_features = 500\n",
42-
"\n",
43-
"learning_rate = 0.001\n",
44-
"algorithm = \"cyclic\"\n",
45-
"random_state=23"
40+
"import cudf\n",
41+
"import numpy as np\n",
42+
"from cuml import make_regression, train_test_split\n",
43+
"from cuml.linear_model import ElasticNet as cuElasticNet, Lasso as cuLasso\n",
44+
"from cuml.metrics.regression import r2_score\n",
45+
"from sklearn.linear_model import ElasticNet as skElasticNet, Lasso as skLasso"
4646
]
4747
},
4848
{
4949
"cell_type": "markdown",
5050
"metadata": {},
5151
"source": [
52-
"## Generate Data"
52+
"## Define Parameters"
5353
]
5454
},
5555
{
@@ -58,9 +58,19 @@
5858
"metadata": {},
5959
"outputs": [],
6060
"source": [
61-
"import cudf\n",
62-
"from cuml import make_regression\n",
63-
"from cuml import train_test_split"
61+
"n_samples = 2**17\n",
62+
"n_features = 500\n",
63+
"\n",
64+
"learning_rate = 0.001\n",
65+
"algorithm = \"cyclic\"\n",
66+
"random_state=23"
67+
]
68+
},
69+
{
70+
"cell_type": "markdown",
71+
"metadata": {},
72+
"source": [
73+
"## Generate Data"
6474
]
6575
},
6676
{
@@ -114,18 +124,15 @@
114124
"metadata": {},
115125
"outputs": [],
116126
"source": [
117-
"import numpy as np\n",
118-
"\n",
119-
"from sklearn.linear_model import Lasso\n",
120-
"\n",
121-
"ols_sk = Lasso(alpha=np.array([learning_rate]),\n",
122-
" fit_intercept = True,\n",
123-
" normalize = False,\n",
124-
" max_iter = 1000,\n",
125-
" selection=algorithm,\n",
126-
" tol=1e-10)\n",
127+
"%%time\n",
128+
"ols_sk = skLasso(alpha=np.array([learning_rate]),\n",
129+
" fit_intercept = True,\n",
130+
" normalize = False,\n",
131+
" max_iter = 1000,\n",
132+
" selection=algorithm,\n",
133+
" tol=1e-10)\n",
127134
"\n",
128-
"%time _ = ols_sk.fit(X_train, y_train)"
135+
"ols_sk.fit(X_train, y_train)"
129136
]
130137
},
131138
{
@@ -144,9 +151,8 @@
144151
"metadata": {},
145152
"outputs": [],
146153
"source": [
147-
"from sklearn.metrics import r2_score\n",
148-
"\n",
149-
"r2_score_sk = r2_score(y_test, predict_sk)"
154+
"%%time\n",
155+
"r2_score_sk = r2_score(y_cudf_test, predict_sk)"
150156
]
151157
},
152158
{
@@ -164,16 +170,15 @@
164170
"metadata": {},
165171
"outputs": [],
166172
"source": [
167-
"from cuml.linear_model import Lasso\n",
168-
"\n",
169-
"ols_cuml = Lasso(alpha=np.array([learning_rate]), \\\n",
170-
" fit_intercept = True, \\\n",
171-
" normalize = False, \\\n",
172-
" max_iter = 1000, \\\n",
173-
" selection=algorithm,\n",
174-
" tol=1e-10)\n",
173+
"%%time\n",
174+
"ols_cuml = cuLasso(alpha=np.array([learning_rate]),\n",
175+
" fit_intercept = True,\n",
176+
" normalize = False,\n",
177+
" max_iter = 1000,\n",
178+
" selection=algorithm,\n",
179+
" tol=1e-10)\n",
175180
"\n",
176-
"%time _ = ols_cuml.fit(X_cudf, y_cudf)"
181+
"ols_cuml.fit(X_cudf, y_cudf)"
177182
]
178183
},
179184
{
@@ -192,8 +197,7 @@
192197
"metadata": {},
193198
"outputs": [],
194199
"source": [
195-
"from cuml.metrics.regression import r2_score\n",
196-
"\n",
200+
"%%time\n",
197201
"r2_score_cuml = r2_score(y_cudf_test, predict_cuml)"
198202
]
199203
},
@@ -236,16 +240,15 @@
236240
"metadata": {},
237241
"outputs": [],
238242
"source": [
239-
"from sklearn.linear_model import ElasticNet\n",
240-
"\n",
241-
"elastic_sk = ElasticNet(alpha=np.array([learning_rate]),\n",
242-
" fit_intercept = True,\n",
243-
" normalize = False,\n",
244-
" max_iter = 1000,\n",
245-
" selection=algorithm,\n",
246-
" tol=1e-10)\n",
243+
"%%time\n",
244+
"elastic_sk = skElasticNet(alpha=np.array([learning_rate]),\n",
245+
" fit_intercept = True,\n",
246+
" normalize = False,\n",
247+
" max_iter = 1000,\n",
248+
" selection=algorithm,\n",
249+
" tol=1e-10)\n",
247250
"\n",
248-
"%time _ = elastic_sk.fit(X_train, y_train)"
251+
"elastic_sk.fit(X_train, y_train)"
249252
]
250253
},
251254
{
@@ -264,9 +267,8 @@
264267
"metadata": {},
265268
"outputs": [],
266269
"source": [
267-
"from sklearn.metrics import r2_score\n",
268-
"\n",
269-
"r2_score_elas_sk = r2_score(y_test, predict_elas_sk)"
270+
"%%time\n",
271+
"r2_score_elas_sk = r2_score(y_cudf_test, predict_elas_sk)"
270272
]
271273
},
272274
{
@@ -284,16 +286,15 @@
284286
"metadata": {},
285287
"outputs": [],
286288
"source": [
287-
"from cuml.linear_model import ElasticNet\n",
288-
"\n",
289-
"elastic_cuml = ElasticNet(alpha=np.array([learning_rate]), \n",
290-
" fit_intercept = True, \n",
291-
" normalize = False, \n",
292-
" max_iter = 1000, \n",
293-
" selection=algorithm, \n",
294-
" tol=1e-10)\n",
289+
"%%time\n",
290+
"elastic_cuml = cuElasticNet(alpha=np.array([learning_rate]), \n",
291+
" fit_intercept = True,\n",
292+
" normalize = False,\n",
293+
" max_iter = 1000,\n",
294+
" selection=algorithm,\n",
295+
" tol=1e-10)\n",
295296
"\n",
296-
"%time _ = elastic_cuml.fit(X_cudf, y_cudf)"
297+
"elastic_cuml.fit(X_cudf, y_cudf)"
297298
]
298299
},
299300
{
@@ -312,8 +313,7 @@
312313
"metadata": {},
313314
"outputs": [],
314315
"source": [
315-
"from cuml.metrics.regression import r2_score\n",
316-
"\n",
316+
"%%time\n",
317317
"r2_score_elas_cuml = r2_score(y_cudf_test, predict_elas_cuml)"
318318
]
319319
},

‎cuml/dbscan_demo.ipynb‎

Lines changed: 36 additions & 37 deletions
Original file line numberDiff line numberDiff line change
@@ -19,7 +19,7 @@
1919
"cell_type": "markdown",
2020
"metadata": {},
2121
"source": [
22-
"## Define Parameters"
22+
"## Imports"
2323
]
2424
},
2525
{
@@ -28,19 +28,23 @@
2828
"metadata": {},
2929
"outputs": [],
3030
"source": [
31-
"n_samples = 10**4\n",
32-
"n_features = 2\n",
31+
"import cudf\n",
32+
"import matplotlib.pyplot as plt\n",
33+
"import numpy as np\n",
3334
"\n",
34-
"eps = 0.15\n",
35-
"min_samples = 3\n",
36-
"random_state = 23"
35+
"from cuml.datasets import make_blobs\n",
36+
"from cuml.cluster import DBSCAN as cuDBSCAN\n",
37+
"from sklearn.cluster import DBSCAN as skDBSCAN\n",
38+
"from sklearn.metrics import adjusted_rand_score\n",
39+
"\n",
40+
"%matplotlib inline"
3741
]
3842
},
3943
{
4044
"cell_type": "markdown",
4145
"metadata": {},
4246
"source": [
43-
"## Generate Data"
47+
"## Define Parameters"
4448
]
4549
},
4650
{
@@ -49,8 +53,19 @@
4953
"metadata": {},
5054
"outputs": [],
5155
"source": [
52-
"import cudf\n",
53-
"from cuml.datasets import make_blobs"
56+
"n_samples = 10**4\n",
57+
"n_features = 2\n",
58+
"\n",
59+
"eps = 0.15\n",
60+
"min_samples = 3\n",
61+
"random_state = 23"
62+
]
63+
},
64+
{
65+
"cell_type": "markdown",
66+
"metadata": {},
67+
"source": [
68+
"## Generate Data"
5469
]
5570
},
5671
{
@@ -97,14 +112,13 @@
97112
"metadata": {},
98113
"outputs": [],
99114
"source": [
100-
"from sklearn.cluster import DBSCAN\n",
101-
"\n",
102-
"clustering_sk = DBSCAN(eps=eps,\n",
103-
" min_samples=min_samples,\n",
104-
" algorithm=\"brute\",\n",
105-
" n_jobs=-1)\n",
115+
"%%time\n",
116+
"clustering_sk = skDBSCAN(eps=eps,\n",
117+
" min_samples=min_samples,\n",
118+
" algorithm=\"brute\",\n",
119+
" n_jobs=-1)\n",
106120
"\n",
107-
"%time _ = clustering_sk.fit(host_data)"
121+
"clustering_sk.fit(host_data)"
108122
]
109123
},
110124
{
@@ -122,14 +136,13 @@
122136
"metadata": {},
123137
"outputs": [],
124138
"source": [
125-
"from cuml.cluster import DBSCAN\n",
126-
"\n",
127-
"clustering_cuml = DBSCAN(eps=eps,\n",
128-
" min_samples=min_samples,\n",
129-
" verbose=True,\n",
130-
" max_mbytes_per_batch=13e3)\n",
139+
"%%time\n",
140+
"clustering_cuml = cuDBSCAN(eps=eps,\n",
141+
" min_samples=min_samples,\n",
142+
" verbose=True,\n",
143+
" max_mbytes_per_batch=13e3)\n",
131144
"\n",
132-
"%time _ = clustering_cuml.fit(device_data, out_dtype=\"int32\")"
145+
"clustering_cuml.fit(device_data, out_dtype=\"int32\")"
133146
]
134147
},
135148
{
@@ -147,11 +160,6 @@
147160
"metadata": {},
148161
"outputs": [],
149162
"source": [
150-
"import matplotlib.pyplot as plt\n",
151-
"import numpy as np\n",
152-
"\n",
153-
"%matplotlib inline\n",
154-
"\n",
155163
"fig = plt.figure(figsize=(16, 10))\n",
156164
"\n",
157165
"X = np.array(host_data)\n",
@@ -187,15 +195,6 @@
187195
"Use the `adjusted_rand_score` to compare the two results, making sure the clusters are labeled similarly by both algorithms even if the exact numerical labels are not identical. "
188196
]
189197
},
190-
{
191-
"cell_type": "code",
192-
"execution_count": null,
193-
"metadata": {},
194-
"outputs": [],
195-
"source": [
196-
"from sklearn.metrics import adjusted_rand_score"
197-
]
198-
},
199198
{
200199
"cell_type": "code",
201200
"execution_count": null,

0 commit comments

Comments
 (0)