[英]Callback with add_loss in Keras
I am trying to implement a VAE style network in Keras.我正在尝试在 Keras 中实现 VAE 风格的网络。 I compute my negative log-likelihood and KL divergence and add them to my model with model.add_loss(NLL)
and model.add_loss(KL)
, respectively.我计算了我的负对数似然和 KL 散度,并分别使用model.add_loss(NLL)
和model.add_loss(KL)
将它们添加到我的模型中。
I would like to scale my KL term as training progresses ("KL-annealing").我想随着训练的进行(“KL 退火”)扩展我的 KL 术语。 I attempted to do this with custom callback ( detailed here ), but the KL loss term does not get updated - the model.add_loss(KL)
term is constant over time, despite the KL weight getting updated (see Figure).我尝试使用自定义回调(此处详细说明)来执行此操作,但 KL 损失项未更新 - 尽管 KL 权重已更新,但model.add_loss(KL)
项随时间保持不变(见图)。
Loss over training epochs How can I make the model.add_loss(KL)
term depend on KL_weight
?训练时期的损失如何使model.add_loss(KL)
项取决于KL_weight
?
Code to demonstrate the idea:代码来演示这个想法:
...
<NLL calculations>
...
# Add the first loss to the model, the NLL:
model.add_loss(NLL)
from keras.callbacks import Callback
klstart = 40
# number of epochs over which KL scaling is increased from 0 to 1
kl_annealtime = 20
weight = tf.keras.backend.variable(0.0) #intialiase KL weight term
class AnnealingCallback(Callback):
def __init__(self, weight):
self.weight = weight
def on_epoch_end (self, epoch, logs={}):
if epoch > klstart :
new_weight = min(tf.keras.backend.get_value(self.weight) + (1./ kl_annealtime), 1.)
tf.keras.backend.set_value(self.weight, new_weight)
print ("Current KL Weight is " + str(tf.keras.backend.get_value(self.weight)))
# Now the KL divergence:
<KL weight calculations>
KL = weight*tf.reduce_mean(tfp.distributions.kl_divergence(p, q))
model.add_loss(KL)
# Now compile the model with a specified optimiser
opt = tf.keras.optimizers.Adam(lr=0.001,clipnorm=0.1)
model.compile(optimizer=opt)
# Monitor how the NLL and KL divergence differ over time
model.add_metric(KL, name='kl_loss', aggregation='mean')
model.add_metric(NLL, name='mse_loss', aggregation='mean')
ops.reset_default_graph()
history=model.fit(Y_portioned, # Input or "Y_true"
verbose=1,
callbacks=[earlystopping_callback,callback_reduce_lr,AnnealingCallback(weight)],
epochs=650,
batch_size=8
) # <- Increase batch size for speed up
Versions: TensorFlow 2.1.0, Keras 2.2.4版本:TensorFlow 2.1.0、Keras 2.2.4
Many thanks in advance提前谢谢了
So I was exactly looking for this implementation in Keras when I came across your post.所以当我看到你的帖子时,我正是在 Keras 中寻找这个实现。 After a while I seem to have found only two mistakes:过了一会儿,我似乎只发现了两个错误:
the corrected version would look like this:更正后的版本如下所示:
#start = klstart
#time = kl_annealtime
class AnnealingCallback(keras.callbacks.Callback):
def __init__(self, weight=tf.keras.backend.variable(0.0), start=20, time=40):
self.weight = weight
self.start = start
self.time = time
def on_epoch_end (self, epoch, logs={}):
if epoch > self.start :
new_weight = min(tf.keras.backend.get_value(self.weight) + (1./self.time), 1.)
tf.keras.backend.set_value(self.weight, new_weight)
print("Current KL Weight is " + str(tf.keras.backend.get_value(self.weight)))
Now you can instantiate the weight:现在您可以实例化权重:
AC = AnnealingCallback()
w = AC.weight
which pre-multpiplies your KL-divergence.它预先乘以您的 KL 散度。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.