2martens/masterthesis
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Reverted to default AdamOptimizer beta values

Browse Source 
Signed-off-by: Jim Martens <github@2martens.de>
This commit is contained in:
Jim Martens
2019-04-12 13:44:05 +02:00
parent 25e2110a5b
commit 7ea63a9a49
1 changed files with 2 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
src/twomartens/masterthesis/aae/train.py
View File

@ -84,8 +84,7 @@ def train_simple(dataset: tf.data.Dataset,
'encoder': model.Encoder(zsize), 'encoder': model.Encoder(zsize),
'decoder': model.Decoder(channels, zsize), 'decoder': model.Decoder(channels, zsize),
# define optimizers # define optimizers
'enc_dec_optimizer': tf.train.AdamOptimizer(learning_rate=checkpointables['learning_rate_var'], 'enc_dec_optimizer': tf.train.AdamOptimizer(learning_rate=checkpointables['learning_rate_var']),
beta1=0.5, beta2=0.999),
# global step counter # global step counter
'epoch_var': K.variable(-1, dtype=tf.int64), 'epoch_var': K.variable(-1, dtype=tf.int64),
'global_step': tf.train.get_or_create_global_step(), 'global_step': tf.train.get_or_create_global_step(),
@ -201,7 +200,7 @@ def _train_enc_dec_step_simple(encoder: model.Encoder, decoder: model.Decoder,
:param encoder: instance of encoder model :param encoder: instance of encoder model
:param decoder: instance of decoder model :param decoder: instance of decoder model
:param optimizer: instance of chosen optimizer :param optimizer: instance of chosen optimizer
:param inputs: inputs from dataset :param inputs: inputs from data set
:param global_step: the global step variable :param global_step: the global step variable
:param global_step_enc_dec: global step variable for enc_dec :param global_step_enc_dec: global step variable for enc_dec
:return: tuple of reconstruction loss, reconstructed input :return: tuple of reconstruction loss, reconstructed input