Skip to content

Commit 5693a88

Browse files
rflamaryrflamary
authored
[MRG] Add github action test cuda (#473)
* add workflow cuda * try with conda * try again conda * and now? * proper python version * cleanup * try to use proper python * again? * cleanup stuff and use python3.10 * remove dr test * remove docstrng tetss for stoc methods * add condition for cuda tests * remove unused files * upate release file * debug workflow
1 parent 7e0ea27 commit 5693a88

File tree

3 files changed

+27
-194
lines changed

3 files changed

+27
-194
lines changed

‎.github/workflows/build_tests_cuda.yml

Lines changed: 25 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,25 @@
1+
name: Tests CUDA
2+
3+
on:
4+
workflow_dispatch:
5+
pull_request_review:
6+
types: [submitted]
7+
push:
8+
branches:
9+
- 'master' # Set a branch to run CI tests on
10+
11+
jobs:
12+
linux-cuda:
13+
14+
runs-on: pc-cuda
15+
if: github.event.review.state == 'approved' || github.event_name == 'workflow_dispatch' || (github.event_name == 'push' && github.event.branch == 'master')
16+
17+
steps:
18+
- uses: actions/checkout@v1
19+
- name: Install POT
20+
run: |
21+
python3.10 -m pip install --ignore-installed -e .
22+
- name: Run tests
23+
run: |
24+
python3.10 -m pytest --durations=20 -v test/ ot/ --doctest-modules --color=yes --ignore=test/test_dr.py --ignore=ot.dr --ignore=ot.plot
25+

‎RELEASES.md

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -5,6 +5,8 @@
55
#### New features
66
- Make alpha parameter in Fused Gromov Wasserstein differentiable (PR #463)
77
- Added the sparsity-constrained OT solver to `ot.smooth` and added ` projection_sparse_simplex` to `ot.utils` (PR #459)
8+
- Add tests on GPU for master branch and approved PR (PR #473)
9+
810
#### Closed issues
911

1012
- Fix circleci-redirector action and codecov (PR #460)

‎ot/stochastic.py

Lines changed: 0 additions & 194 deletions
Original file line numberDiff line numberDiff line change
@@ -58,27 +58,6 @@ def coordinate_grad_semi_dual(b, M, reg, beta, i):
5858
-------
5959
coordinate gradient : ndarray, shape (nt,)
6060
61-
Examples
62-
--------
63-
>>> import ot
64-
>>> np.random.seed(0)
65-
>>> n_source = 7
66-
>>> n_target = 4
67-
>>> a = ot.utils.unif(n_source)
68-
>>> b = ot.utils.unif(n_target)
69-
>>> X_source = np.random.randn(n_source, 2)
70-
>>> Y_target = np.random.randn(n_target, 2)
71-
>>> M = ot.dist(X_source, Y_target)
72-
>>> ot.stochastic.solve_semi_dual_entropic(a, b, M, reg=1, method="ASGD", numItermax=300000)
73-
array([[2.53942342e-02, 9.98640673e-02, 1.75945647e-02, 4.27664307e-06],
74-
[1.21556999e-01, 1.26350515e-02, 1.30491795e-03, 7.36017394e-03],
75-
[3.54070702e-03, 7.63581358e-02, 6.29581672e-02, 1.32812798e-07],
76-
[2.60578198e-02, 3.35916645e-02, 8.28023223e-02, 4.05336238e-04],
77-
[9.86808864e-03, 7.59774324e-04, 1.08702729e-02, 1.21359007e-01],
78-
[2.17218856e-02, 9.12931802e-04, 1.87962526e-03, 1.18342700e-01],
79-
[4.14237512e-02, 2.67487857e-02, 7.23016955e-02, 2.38291052e-03]])
80-
81-
8261
.. _references-coordinate-grad-semi-dual:
8362
References
8463
----------
@@ -137,27 +116,6 @@ def sag_entropic_transport(a, b, M, reg, numItermax=10000, lr=None):
137116
v : ndarray, shape (`nt`,)
138117
Dual variable.
139118
140-
Examples
141-
--------
142-
>>> import ot
143-
>>> np.random.seed(0)
144-
>>> n_source = 7
145-
>>> n_target = 4
146-
>>> a = ot.utils.unif(n_source)
147-
>>> b = ot.utils.unif(n_target)
148-
>>> X_source = np.random.randn(n_source, 2)
149-
>>> Y_target = np.random.randn(n_target, 2)
150-
>>> M = ot.dist(X_source, Y_target)
151-
>>> ot.stochastic.solve_semi_dual_entropic(a, b, M, reg=1, method="ASGD", numItermax=300000)
152-
array([[2.53942342e-02, 9.98640673e-02, 1.75945647e-02, 4.27664307e-06],
153-
[1.21556999e-01, 1.26350515e-02, 1.30491795e-03, 7.36017394e-03],
154-
[3.54070702e-03, 7.63581358e-02, 6.29581672e-02, 1.32812798e-07],
155-
[2.60578198e-02, 3.35916645e-02, 8.28023223e-02, 4.05336238e-04],
156-
[9.86808864e-03, 7.59774324e-04, 1.08702729e-02, 1.21359007e-01],
157-
[2.17218856e-02, 9.12931802e-04, 1.87962526e-03, 1.18342700e-01],
158-
[4.14237512e-02, 2.67487857e-02, 7.23016955e-02, 2.38291052e-03]])
159-
160-
161119
.. _references-sag-entropic-transport:
162120
References
163121
----------
@@ -225,27 +183,6 @@ def averaged_sgd_entropic_transport(a, b, M, reg, numItermax=300000, lr=None):
225183
ave_v : ndarray, shape (`nt`,)
226184
dual variable
227185
228-
Examples
229-
--------
230-
>>> import ot
231-
>>> np.random.seed(0)
232-
>>> n_source = 7
233-
>>> n_target = 4
234-
>>> a = ot.utils.unif(n_source)
235-
>>> b = ot.utils.unif(n_target)
236-
>>> X_source = np.random.randn(n_source, 2)
237-
>>> Y_target = np.random.randn(n_target, 2)
238-
>>> M = ot.dist(X_source, Y_target)
239-
>>> ot.stochastic.solve_semi_dual_entropic(a, b, M, reg=1, method="ASGD", numItermax=300000)
240-
array([[2.53942342e-02, 9.98640673e-02, 1.75945647e-02, 4.27664307e-06],
241-
[1.21556999e-01, 1.26350515e-02, 1.30491795e-03, 7.36017394e-03],
242-
[3.54070702e-03, 7.63581358e-02, 6.29581672e-02, 1.32812798e-07],
243-
[2.60578198e-02, 3.35916645e-02, 8.28023223e-02, 4.05336238e-04],
244-
[9.86808864e-03, 7.59774324e-04, 1.08702729e-02, 1.21359007e-01],
245-
[2.17218856e-02, 9.12931802e-04, 1.87962526e-03, 1.18342700e-01],
246-
[4.14237512e-02, 2.67487857e-02, 7.23016955e-02, 2.38291052e-03]])
247-
248-
249186
.. _references-averaged-sgd-entropic-transport:
250187
References
251188
----------
@@ -304,27 +241,6 @@ def c_transform_entropic(b, M, reg, beta):
304241
u : ndarray, shape (`ns`,)
305242
Dual variable.
306243
307-
Examples
308-
--------
309-
>>> import ot
310-
>>> np.random.seed(0)
311-
>>> n_source = 7
312-
>>> n_target = 4
313-
>>> a = ot.utils.unif(n_source)
314-
>>> b = ot.utils.unif(n_target)
315-
>>> X_source = np.random.randn(n_source, 2)
316-
>>> Y_target = np.random.randn(n_target, 2)
317-
>>> M = ot.dist(X_source, Y_target)
318-
>>> ot.stochastic.solve_semi_dual_entropic(a, b, M, reg=1, method="ASGD", numItermax=300000)
319-
array([[2.53942342e-02, 9.98640673e-02, 1.75945647e-02, 4.27664307e-06],
320-
[1.21556999e-01, 1.26350515e-02, 1.30491795e-03, 7.36017394e-03],
321-
[3.54070702e-03, 7.63581358e-02, 6.29581672e-02, 1.32812798e-07],
322-
[2.60578198e-02, 3.35916645e-02, 8.28023223e-02, 4.05336238e-04],
323-
[9.86808864e-03, 7.59774324e-04, 1.08702729e-02, 1.21359007e-01],
324-
[2.17218856e-02, 9.12931802e-04, 1.87962526e-03, 1.18342700e-01],
325-
[4.14237512e-02, 2.67487857e-02, 7.23016955e-02, 2.38291052e-03]])
326-
327-
328244
.. _references-c-transform-entropic:
329245
References
330246
----------
@@ -399,27 +315,6 @@ def solve_semi_dual_entropic(a, b, M, reg, method, numItermax=10000, lr=None,
399315
log : dict
400316
log dictionary return only if log==True in parameters
401317
402-
Examples
403-
--------
404-
>>> import ot
405-
>>> np.random.seed(0)
406-
>>> n_source = 7
407-
>>> n_target = 4
408-
>>> a = ot.utils.unif(n_source)
409-
>>> b = ot.utils.unif(n_target)
410-
>>> X_source = np.random.randn(n_source, 2)
411-
>>> Y_target = np.random.randn(n_target, 2)
412-
>>> M = ot.dist(X_source, Y_target)
413-
>>> ot.stochastic.solve_semi_dual_entropic(a, b, M, reg=1, method="ASGD", numItermax=300000)
414-
array([[2.53942342e-02, 9.98640673e-02, 1.75945647e-02, 4.27664307e-06],
415-
[1.21556999e-01, 1.26350515e-02, 1.30491795e-03, 7.36017394e-03],
416-
[3.54070702e-03, 7.63581358e-02, 6.29581672e-02, 1.32812798e-07],
417-
[2.60578198e-02, 3.35916645e-02, 8.28023223e-02, 4.05336238e-04],
418-
[9.86808864e-03, 7.59774324e-04, 1.08702729e-02, 1.21359007e-01],
419-
[2.17218856e-02, 9.12931802e-04, 1.87962526e-03, 1.18342700e-01],
420-
[4.14237512e-02, 2.67487857e-02, 7.23016955e-02, 2.38291052e-03]])
421-
422-
423318
.. _references-solve-semi-dual-entropic:
424319
References
425320
----------
@@ -509,33 +404,6 @@ def batch_grad_dual(a, b, M, reg, alpha, beta, batch_size, batch_alpha,
509404
grad : ndarray, shape (`ns`,)
510405
partial grad F
511406
512-
Examples
513-
--------
514-
>>> import ot
515-
>>> np.random.seed(0)
516-
>>> n_source = 7
517-
>>> n_target = 4
518-
>>> a = ot.utils.unif(n_source)
519-
>>> b = ot.utils.unif(n_target)
520-
>>> X_source = np.random.randn(n_source, 2)
521-
>>> Y_target = np.random.randn(n_target, 2)
522-
>>> M = ot.dist(X_source, Y_target)
523-
>>> sgd_dual_pi, log = ot.stochastic.solve_dual_entropic(a, b, M, reg=1, batch_size=3, numItermax=30000, lr=0.1, log=True)
524-
>>> log['alpha']
525-
array([0.71759102, 1.57057384, 0.85576566, 0.1208211 , 0.59190466,
526-
1.197148 , 0.17805133])
527-
>>> log['beta']
528-
array([0.49741367, 0.57478564, 1.40075528, 2.75890102])
529-
>>> sgd_dual_pi
530-
array([[2.09730063e-02, 8.38169324e-02, 7.50365455e-03, 8.72731415e-09],
531-
[5.58432437e-03, 5.89881299e-04, 3.09558411e-05, 8.35469849e-07],
532-
[3.26489515e-03, 7.15536035e-02, 2.99778211e-02, 3.02601593e-10],
533-
[4.05390622e-02, 5.31085068e-02, 6.65191787e-02, 1.55812785e-06],
534-
[7.82299812e-02, 6.12099102e-03, 4.44989098e-02, 2.37719187e-03],
535-
[5.06266486e-02, 2.16230494e-03, 2.26215141e-03, 6.81514609e-04],
536-
[6.06713990e-02, 3.98139808e-02, 5.46829338e-02, 8.62371424e-06]])
537-
538-
539407
.. _references-batch-grad-dual:
540408
References
541409
----------
@@ -600,37 +468,6 @@ def sgd_entropic_regularization(a, b, M, reg, batch_size, numItermax, lr):
600468
beta : ndarray, shape (nt,)
601469
dual variable
602470
603-
Examples
604-
--------
605-
>>> import ot
606-
>>> n_source = 7
607-
>>> n_target = 4
608-
>>> reg = 1
609-
>>> numItermax = 20000
610-
>>> lr = 0.1
611-
>>> batch_size = 3
612-
>>> log = True
613-
>>> a = ot.utils.unif(n_source)
614-
>>> b = ot.utils.unif(n_target)
615-
>>> rng = np.random.RandomState(0)
616-
>>> X_source = rng.randn(n_source, 2)
617-
>>> Y_target = rng.randn(n_target, 2)
618-
>>> M = ot.dist(X_source, Y_target)
619-
>>> sgd_dual_pi, log = ot.stochastic.solve_dual_entropic(a, b, M, reg, batch_size, numItermax, lr, log)
620-
>>> log['alpha']
621-
array([0.64171798, 1.27932201, 0.78132257, 0.15638935, 0.54888354,
622-
1.03663469, 0.20595781])
623-
>>> log['beta']
624-
array([0.51207194, 0.58033189, 1.28922676, 2.26859736])
625-
>>> sgd_dual_pi
626-
array([[1.97276541e-02, 7.81248547e-02, 6.22136048e-03, 4.95442423e-09],
627-
[4.23494310e-03, 4.43286263e-04, 2.06927079e-05, 3.82389139e-07],
628-
[3.07542414e-03, 6.67897769e-02, 2.48904999e-02, 1.72030247e-10],
629-
[4.26271990e-02, 5.53375455e-02, 6.16535024e-02, 9.88812650e-07],
630-
[7.60423265e-02, 5.89585256e-03, 3.81267087e-02, 1.39458256e-03],
631-
[4.37557504e-02, 1.85189176e-03, 1.72335760e-03, 3.55491279e-04],
632-
[6.33096109e-02, 4.11683954e-02, 5.02962051e-02, 5.43097516e-06]])
633-
634471
References
635472
----------
636473
.. [19] Seguy, V., Bhushan Damodaran, B., Flamary, R., Courty, N., Rolet, A.& Blondel, M. Large-scale Optimal Transport and Mapping Estimation. International Conference on Learning Representation (2018)
@@ -702,37 +539,6 @@ def solve_dual_entropic(a, b, M, reg, batch_size, numItermax=10000, lr=1,
702539
log : dict
703540
log dictionary return only if log==True in parameters
704541
705-
Examples
706-
--------
707-
>>> import ot
708-
>>> n_source = 7
709-
>>> n_target = 4
710-
>>> reg = 1
711-
>>> numItermax = 20000
712-
>>> lr = 0.1
713-
>>> batch_size = 3
714-
>>> log = True
715-
>>> a = ot.utils.unif(n_source)
716-
>>> b = ot.utils.unif(n_target)
717-
>>> rng = np.random.RandomState(0)
718-
>>> X_source = rng.randn(n_source, 2)
719-
>>> Y_target = rng.randn(n_target, 2)
720-
>>> M = ot.dist(X_source, Y_target)
721-
>>> sgd_dual_pi, log = ot.stochastic.solve_dual_entropic(a, b, M, reg, batch_size, numItermax, lr, log)
722-
>>> log['alpha']
723-
array([0.64057733, 1.2683513 , 0.75610161, 0.16024284, 0.54926534,
724-
1.0514201 , 0.19958936])
725-
>>> log['beta']
726-
array([0.51372571, 0.58843489, 1.27993921, 2.24344807])
727-
>>> sgd_dual_pi
728-
array([[1.97377795e-02, 7.86706853e-02, 6.15682001e-03, 4.82586997e-09],
729-
[4.19566963e-03, 4.42016865e-04, 2.02777272e-05, 3.68823708e-07],
730-
[3.00379244e-03, 6.56562018e-02, 2.40462171e-02, 1.63579656e-10],
731-
[4.28626062e-02, 5.60031599e-02, 6.13193826e-02, 9.67977735e-07],
732-
[7.61972739e-02, 5.94609051e-03, 3.77886693e-02, 1.36046648e-03],
733-
[4.44810042e-02, 1.89476742e-03, 1.73285847e-03, 3.51826036e-04],
734-
[6.30118293e-02, 4.12398660e-02, 4.95148998e-02, 5.26247246e-06]])
735-
736542
References
737543
----------
738544
.. [19] Seguy, V., Bhushan Damodaran, B., Flamary, R., Courty, N., Rolet, A.& Blondel, M. Large-scale Optimal Transport and Mapping Estimation. International Conference on Learning Representation (2018)

0 commit comments

Comments
 (0)