From 7b86681defc3ca2182d8bbdd7ce6d14f982c747b Mon Sep 17 00:00:00 2001
From: lucas_miranda <lucasmiranda42@gmail.com>
Date: Sun, 7 Jun 2020 13:02:12 +0200
Subject: [PATCH] Instanciated a mean and a variance for each component, and
the categorical prior as a Dense layer with a softmax activation
---
source/models.py | 56 +++++++++++++++++++++++++-----------------------
1 file changed, 29 insertions(+), 27 deletions(-)
diff --git a/source/models.py b/source/models.py
index c21bbcfc..cfb9d626 100644
--- a/source/models.py
+++ b/source/models.py
@@ -344,9 +344,7 @@ class SEQ_2_SEQ_VAE:
return self.input_shape[1:] * huber(x_, x_decoded_mean_)
vae.compile(
- loss=huber_loss,
- optimizer=Adam(lr=self.learn_rate,),
- metrics=["mae"],
+ loss=huber_loss, optimizer=Adam(lr=self.learn_rate,), metrics=["mae"],
)
return encoder, generator, vae, kl_warmup_callback, mmd_warmup_callback
@@ -587,9 +585,7 @@ class SEQ_2_SEQ_VAEP:
return self.input_shape[1:] * huber(x_, x_decoded_mean_)
vaep.compile(
- loss=huber_loss,
- optimizer=Adam(lr=self.learn_rate,),
- metrics=["mae"],
+ loss=huber_loss, optimizer=Adam(lr=self.learn_rate,), metrics=["mae"],
)
return encoder, generator, vaep, kl_warmup_callback, mmd_warmup_callback
@@ -597,19 +593,19 @@ class SEQ_2_SEQ_VAEP:
class SEQ_2_SEQ_MMVAEP:
def __init__(
- self,
- input_shape,
- CONV_filters=256,
- LSTM_units_1=256,
- LSTM_units_2=64,
- DENSE_2=64,
- DROPOUT_RATE=0.25,
- ENCODING=32,
- learn_rate=1e-3,
- loss="ELBO+MMD",
- kl_warmup_epochs=0,
- mmd_warmup_epochs=0,
- number_of_components=1,
+ self,
+ input_shape,
+ CONV_filters=256,
+ LSTM_units_1=256,
+ LSTM_units_2=64,
+ DENSE_2=64,
+ DROPOUT_RATE=0.25,
+ ENCODING=32,
+ learn_rate=1e-3,
+ loss="ELBO+MMD",
+ kl_warmup_epochs=0,
+ mmd_warmup_epochs=0,
+ number_of_components=1,
):
self.input_shape = input_shape
self.CONV_filters = CONV_filters
@@ -626,7 +622,11 @@ class SEQ_2_SEQ_MMVAEP:
self.number_of_components = number_of_components
assert (
- "ELBO" in self.loss or "MMD" in self.loss
+ self.number_of_components > 0
+ ), "The number of components must be an integer greater than zero"
+
+ assert (
+ "ELBO" in self.loss or "MMD" in self.loss
), "loss must be one of ELBO, MMD or ELBO+MMD (default)"
def build(self):
@@ -684,8 +684,8 @@ class SEQ_2_SEQ_MMVAEP:
Model_D0 = DenseTranspose(
Model_E5, activation="relu", output_dim=self.ENCODING,
)
- Model_D1 = DenseTranspose(Model_E4, activation="relu", output_dim=self.DENSE_2, )
- Model_D2 = DenseTranspose(Model_E3, activation="relu", output_dim=self.DENSE_1, )
+ Model_D1 = DenseTranspose(Model_E4, activation="relu", output_dim=self.DENSE_2,)
+ Model_D2 = DenseTranspose(Model_E3, activation="relu", output_dim=self.DENSE_1,)
Model_D3 = RepeatVector(self.input_shape[1])
Model_D4 = Bidirectional(
LSTM(
@@ -719,8 +719,12 @@ class SEQ_2_SEQ_MMVAEP:
encoder = BatchNormalization()(encoder)
encoder = Model_E5(encoder)
- z_mean = Dense(self.ENCODING)(encoder)
- z_log_sigma = Dense(self.ENCODING)(encoder)
+ # Categorical prior on mixture of Gaussians
+ categories = Dense(self.number_of_components, activation="softmax")
+
+ for _ in range(self.number_of_components):
+ z_mean = Dense(self.ENCODING)(encoder)
+ z_log_sigma = Dense(self.ENCODING)(encoder)
# Define and control custom loss functions
kl_warmup_callback = False
@@ -832,9 +836,7 @@ class SEQ_2_SEQ_MMVAEP:
return self.input_shape[1:] * huber(x_, x_decoded_mean_)
gmvaep.compile(
- loss=huber_loss,
- optimizer=Adam(lr=self.learn_rate, ),
- metrics=["mae"],
+ loss=huber_loss, optimizer=Adam(lr=self.learn_rate,), metrics=["mae"],
)
return encoder, generator, gmvaep, kl_warmup_callback, mmd_warmup_callback
--
GitLab