experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Decrease tolerances on ScaleTriL tests to prevent flakiness.

Browse Source 
PiperOrigin-RevId: 246044331
This commit is contained in:
A. Unique TensorFlower 2019-04-30 16:24:23 -07:00 committed by TensorFlower Gardener
parent a198279b08
commit b962b1f193
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
tensorflow/contrib/distributions/python/kernel_tests/bijectors/scale_tril_test.py
View File

@ -41,10 +41,10 @@ class ScaleTriLBijectorTest(test.TestCase):
diag_shift=shift) diag_shift=shift)
y_ = self.evaluate(b.forward(x)) y_ = self.evaluate(b.forward(x))
self.assertAllClose(y, y_) self.assertAllClose(y, y_, rtol=1e-4)
x_ = self.evaluate(b.inverse(y)) x_ = self.evaluate(b.inverse(y))
self.assertAllClose(x, x_) self.assertAllClose(x, x_, rtol=1e-4)
@test_util.run_in_graph_and_eager_modes @test_util.run_in_graph_and_eager_modes
def testInvertible(self): def testInvertible(self):
@ -52,18 +52,18 @@ class ScaleTriLBijectorTest(test.TestCase):
# Generate random inputs from an unconstrained space, with # Generate random inputs from an unconstrained space, with
# event size 6 to specify 3x3 triangular matrices. # event size 6 to specify 3x3 triangular matrices.
batch_shape = [2, 1] batch_shape = [2, 1]
x = np.float32(np.random.randn(*(batch_shape + [6]))) x = np.float32(self._rng.randn(*(batch_shape + [6])))
b = bijectors.ScaleTriL(diag_bijector=bijectors.Softplus(), b = bijectors.ScaleTriL(diag_bijector=bijectors.Softplus(),
diag_shift=3.14159) diag_shift=3.14159)
y = self.evaluate(b.forward(x)) y = self.evaluate(b.forward(x))
self.assertAllEqual(y.shape, batch_shape + [3, 3]) self.assertAllEqual(y.shape, batch_shape + [3, 3])
x_ = self.evaluate(b.inverse(y)) x_ = self.evaluate(b.inverse(y))
self.assertAllClose(x, x_) self.assertAllClose(x, x_, rtol=1e-4)
fldj = self.evaluate(b.forward_log_det_jacobian(x, event_ndims=1)) fldj = self.evaluate(b.forward_log_det_jacobian(x, event_ndims=1))
ildj = self.evaluate(b.inverse_log_det_jacobian(y, event_ndims=2)) ildj = self.evaluate(b.inverse_log_det_jacobian(y, event_ndims=2))
self.assertAllClose(fldj, -ildj) self.assertAllClose(fldj, -ildj, rtol=1e-4)
if __name__ == "__main__": if __name__ == "__main__":
test.main() test.main()