ruby/tensorflow_privacy
1
0
Fork 0
forked from 626_privacy/tensorflow_privacy
Code Pull requests Activity

Improve numerical stability of RDP computation.

Browse source 
PiperOrigin-RevId: 628482307
This commit is contained in:
Galen Andrew 2024-04-26 12:40:13 -07:00 committed by A. Unique TensorFlower
parent 3deaae30a1
commit 8b3d8e132f
1 changed files with 4 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
tensorflow_privacy/privacy/analysis/compute_dp_sgd_privacy_test.py
View file

@ -40,8 +40,8 @@ class ComputeDpSgdPrivacyTest(parameterized.TestCase):
@parameterized.named_parameters(
('Test0', 60000, 150, 1.3, 15, 1e-5, 0.7242234026109595, 19.0),
('Test1', 100000, 100, 1.0, 30, 1e-7, 1.4154988495444845, 13.0),
('Test2', 100000000, 1024, 0.1, 10, 1e-7, 5907982.31138195, 1.25),
('Test1', 100000, 100, 1.0, 30, 1e-7, 1.4154988495444842, 13.0),
('Test2', 100000000, 1024, 0.1, 10, 1e-7, 5907982.3113984, 1.25),
)
def test_compute_dp_sgd_privacy(self, n, batch_size, noise_multiplier, epochs,
delta, expected_eps, expected_order):
@ -85,8 +85,8 @@ class ComputeDpSgdPrivacyTest(parameterized.TestCase):
@parameterized.named_parameters(
('no_microbatching_no_subsampling_rdp', False, None, _RDP, 10.8602036),
('microbatching_no_subsampling_rdp', True, None, _RDP, 26.2880374),
('no_microbatching_with_subsampling_rdp', False, 1e-2, _RDP, 3.2391922),
('microbatching_with_subsampling_rdp', True, 1e-2, _RDP, 22.5970358),
('no_microbatching_with_subsampling_rdp', False, 1e-2, _RDP, 3.2392676),
('microbatching_with_subsampling_rdp', True, 1e-2, _RDP, 22.6248601),
('no_microbatching_no_subsampling_pld', False, None, _PLD, 10.1224946),
('microbatching_no_subsampling_pld', True, None, _PLD, 24.7160779),
('no_microbatching_with_subsampling_pld', False, 1e-2, _PLD, 2.4612381),