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

added a ledger during optimizer instantiation to the language model tutorial

Browse source
This commit is contained in:
Tim Garnsey 2019-04-17 12:52:08 +10:00
parent 51e29667d9
commit aeb6a94b59
1 changed files with 11 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
tutorials/lm_dpsgd_tutorial.py
View file

@ -40,13 +40,14 @@ import numpy as np
import tensorflow as tf import tensorflow as tf
import tensorflow_datasets as tfds import tensorflow_datasets as tfds
from privacy.analysis import privacy_ledger
from privacy.analysis.rdp_accountant import compute_rdp from privacy.analysis.rdp_accountant import compute_rdp
from privacy.analysis.rdp_accountant import get_privacy_spent from privacy.analysis.rdp_accountant import get_privacy_spent
from privacy.optimizers import dp_optimizer from privacy.optimizers import dp_optimizer
tf.flags.DEFINE_boolean('dpsgd', True, 'If True, train with DP-SGD. If False, ' tf.flags.DEFINE_boolean('dpsgd', True, 'If True, train with DP-SGD. If False, '
'train with vanilla SGD.') 'train with vanilla SGD.')
tf.flags.DEFINE_float('learning_rate', .001, 'Learning rate for training') tf.flags.DEFINE_float('learning_rate', 0.001, 'Learning rate for training')
tf.flags.DEFINE_float('noise_multiplier', 0.001, tf.flags.DEFINE_float('noise_multiplier', 0.001,
'Ratio of the standard deviation to the clipping norm') 'Ratio of the standard deviation to the clipping norm')
tf.flags.DEFINE_float('l2_norm_clip', 1.0, 'Clipping norm') tf.flags.DEFINE_float('l2_norm_clip', 1.0, 'Clipping norm')
@ -84,13 +85,20 @@ def rnn_model_fn(features, labels, mode): # pylint: disable=unused-argument
# Configure the training op (for TRAIN mode). # Configure the training op (for TRAIN mode).
if mode == tf.estimator.ModeKeys.TRAIN: if mode == tf.estimator.ModeKeys.TRAIN:
if FLAGS.dpsgd: if FLAGS.dpsgd:
ledger = privacy_ledger.PrivacyLedger(
population_size=NB_TRAIN,
selection_probability=(FLAGS.batch_size / NB_TRAIN),
max_samples=1e6,
max_queries=1e6)
optimizer = dp_optimizer.DPAdamGaussianOptimizer( optimizer = dp_optimizer.DPAdamGaussianOptimizer(
l2_norm_clip=FLAGS.l2_norm_clip, l2_norm_clip=FLAGS.l2_norm_clip,
noise_multiplier=FLAGS.noise_multiplier, noise_multiplier=FLAGS.noise_multiplier,
num_microbatches=FLAGS.microbatches, num_microbatches=FLAGS.microbatches,
ledger=ledger,
learning_rate=FLAGS.learning_rate, learning_rate=FLAGS.learning_rate,
unroll_microbatches=True, unroll_microbatches=True)
population_size=NB_TRAIN)
opt_loss = vector_loss opt_loss = vector_loss
else: else:
optimizer = tf.train.AdamOptimizer( optimizer = tf.train.AdamOptimizer(