steffen/ae_toolbox_torch
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Done: Latent Space Viz

Browse Source 
ToDo: Visualization for variational spaces
Trajectory Coloring
Post Processing
Metric
Slurm Skript
This commit is contained in:
Si11ium 2019-08-23 09:54:00 +02:00
parent 744c0c50b7
commit 1a0400d736
9 changed files with 159 additions and 76 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

67
.idea/workspace.xml generated
View File

@ -2,14 +2,15 @@
<project version="4"> <project version="4">
<component name="ChangeListManager"> <component name="ChangeListManager">
<list default="true" id="5955480a-c876-43d5-afd7-8717f51f413e" name="Default Changelist" comment=""> <list default="true" id="5955480a-c876-43d5-afd7-8717f51f413e" name="Default Changelist" comment="">
<change afterPath="$PROJECT_DIR$/networks/seperating_adversarial_auto_encoder.py" afterDir="false" />
<change afterPath="$PROJECT_DIR$/viz/viz_latent.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/.idea/workspace.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/workspace.xml" afterDir="false" /> <change beforePath="$PROJECT_DIR$/.idea/workspace.xml" beforeDir="false" afterPath="$PROJECT_DIR$/.idea/workspace.xml" afterDir="false" />
<change beforePath="$PROJECT_DIR$/dataset.py" beforeDir="false" afterPath="$PROJECT_DIR$/dataset.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/networks/adverserial_auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/adverserial_auto_encoder.py" afterDir="false" /> <change beforePath="$PROJECT_DIR$/networks/adverserial_auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/adverserial_auto_encoder.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/networks/auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/auto_encoder.py" afterDir="false" /> <change beforePath="$PROJECT_DIR$/networks/auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/auto_encoder.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/networks/modules.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/modules.py" afterDir="false" /> <change beforePath="$PROJECT_DIR$/networks/modules.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/modules.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/networks/seperating_adversarial_auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/seperating_adversarial_auto_encoder.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/networks/variational_auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/variational_auto_encoder.py" afterDir="false" /> <change beforePath="$PROJECT_DIR$/networks/variational_auto_encoder.py" beforeDir="false" afterPath="$PROJECT_DIR$/networks/variational_auto_encoder.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/run_models.py" beforeDir="false" afterPath="$PROJECT_DIR$/run_models.py" afterDir="false" /> <change beforePath="$PROJECT_DIR$/run_models.py" beforeDir="false" afterPath="$PROJECT_DIR$/run_models.py" afterDir="false" />
<change beforePath="$PROJECT_DIR$/viz/viz_latent.py" beforeDir="false" afterPath="$PROJECT_DIR$/viz/viz_latent.py" afterDir="false" />
</list> </list>
<option name="EXCLUDED_CONVERTED_TO_IGNORED" value="true" /> <option name="EXCLUDED_CONVERTED_TO_IGNORED" value="true" />
<option name="SHOW_DIALOG" value="false" /> <option name="SHOW_DIALOG" value="false" />
@ -68,7 +69,7 @@
</list> </list>
</option> </option>
</component> </component>
<component name="RunManager" selected="Python.run_models"> <component name="RunManager" selected="Python.viz_latent">
<configuration default="true" type="PythonConfigurationType" factoryName="Python"> <configuration default="true" type="PythonConfigurationType" factoryName="Python">
<module name="ae_toolbox_torch" /> <module name="ae_toolbox_torch" />
<option name="INTERPRETER_OPTIONS" value="" /> <option name="INTERPRETER_OPTIONS" value="" />
@ -95,6 +96,9 @@
<module name="ae_toolbox_torch" /> <module name="ae_toolbox_torch" />
<option name="INTERPRETER_OPTIONS" value="" /> <option name="INTERPRETER_OPTIONS" value="" />
<option name="PARENT_ENVS" value="true" /> <option name="PARENT_ENVS" value="true" />
<envs>
<env name="PYTHONUNBUFFERED" value="1" />
</envs>
<option name="SDK_HOME" value="" /> <option name="SDK_HOME" value="" />
<option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" /> <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />
<option name="IS_MODULE_SDK" value="true" /> <option name="IS_MODULE_SDK" value="true" />
@ -129,12 +133,33 @@
<option name="INPUT_FILE" value="" /> <option name="INPUT_FILE" value="" />
<method v="2" /> <method v="2" />
</configuration> </configuration>
<configuration name="viz_latent" type="PythonConfigurationType" factoryName="Python" temporary="true">
<module name="ae_toolbox_torch" />
<option name="INTERPRETER_OPTIONS" value="" />
<option name="PARENT_ENVS" value="true" />
<option name="SDK_HOME" value="" />
<option name="WORKING_DIRECTORY" value="$PROJECT_DIR$/viz" />
<option name="IS_MODULE_SDK" value="true" />
<option name="ADD_CONTENT_ROOTS" value="true" />
<option name="ADD_SOURCE_ROOTS" value="true" />
<EXTENSION ID="PythonCoverageRunConfigurationExtension" runner="coverage.py" />
<option name="SCRIPT_NAME" value="$PROJECT_DIR$/viz/viz_latent.py" />
<option name="PARAMETERS" value="" />
<option name="SHOW_COMMAND_LINE" value="true" />
<option name="EMULATE_TERMINAL" value="false" />
<option name="MODULE_MODE" value="false" />
<option name="REDIRECT_INPUT" value="false" />
<option name="INPUT_FILE" value="" />
<method v="2" />
</configuration>
<list> <list>
<item itemvalue="Python.run_basic_ae" /> <item itemvalue="Python.run_basic_ae" />
<item itemvalue="Python.run_models" /> <item itemvalue="Python.run_models" />
<item itemvalue="Python.viz_latent" />
</list> </list>
<recent_temporary> <recent_temporary>
<list> <list>
<item itemvalue="Python.viz_latent" />
<item itemvalue="Python.run_models" /> <item itemvalue="Python.run_models" />
<item itemvalue="Python.run_basic_ae" /> <item itemvalue="Python.run_basic_ae" />
</list> </list>
@ -153,7 +178,8 @@
<workItem from="1564587420277" duration="6891000" /> <workItem from="1564587420277" duration="6891000" />
<workItem from="1565364574595" duration="1092000" /> <workItem from="1565364574595" duration="1092000" />
<workItem from="1565592214301" duration="53660000" /> <workItem from="1565592214301" duration="53660000" />
<workItem from="1565793671730" duration="53296000" /> <workItem from="1565793671730" duration="53351000" />
<workItem from="1566372837067" duration="27899000" />
</task> </task>
<task id="LOCAL-00001" summary="Lightning integration basic ae, dataloaders and dataset"> <task id="LOCAL-00001" summary="Lightning integration basic ae, dataloaders and dataset">
<created>1565793753423</created> <created>1565793753423</created>
@ -183,7 +209,14 @@
<option name="project" value="LOCAL" /> <option name="project" value="LOCAL" />
<updated>1566064016196</updated> <updated>1566064016196</updated>
</task> </task>
<option name="localTasksCounter" value="5" /> <task id="LOCAL-00005" summary="Done: First VIsualization&#10;ToDo: Visualization for all classes, latent space setups">
<created>1566366992088</created>
<option name="number" value="00005" />
<option name="presentableId" value="LOCAL-00005" />
<option name="project" value="LOCAL" />
<updated>1566366992088</updated>
</task>
<option name="localTasksCounter" value="6" />
<servers /> <servers />
</component> </component>
<component name="TypeScriptGeneratedFilesManager"> <component name="TypeScriptGeneratedFilesManager">
@ -205,32 +238,46 @@
<component name="VcsManagerConfiguration"> <component name="VcsManagerConfiguration">
<MESSAGE value="Lightning integration basic ae, dataloaders and dataset" /> <MESSAGE value="Lightning integration basic ae, dataloaders and dataset" />
<MESSAGE value="Done: AE, VAE, AAE&#10;ToDo: Double AAE, Visualization&#10;All Modularized" /> <MESSAGE value="Done: AE, VAE, AAE&#10;ToDo: Double AAE, Visualization&#10;All Modularized" />
<option name="LAST_COMMIT_MESSAGE" value="Done: AE, VAE, AAE&#10;ToDo: Double AAE, Visualization&#10;All Modularized" /> <MESSAGE value="Done: First VIsualization&#10;ToDo: Visualization for all classes, latent space setups" />
<option name="LAST_COMMIT_MESSAGE" value="Done: First VIsualization&#10;ToDo: Visualization for all classes, latent space setups" />
</component> </component>
<component name="XDebuggerManager"> <component name="XDebuggerManager">
<breakpoint-manager> <breakpoint-manager>
<breakpoints> <breakpoints>
<line-breakpoint enabled="true" suspend="THREAD" type="python-line"> <line-breakpoint enabled="true" suspend="THREAD" type="python-line">
<url>file://$PROJECT_DIR$/run_models.py</url> <url>file://$PROJECT_DIR$/networks/modules.py</url>
<line>206</line>
<option name="timeStamp" value="51" />
</line-breakpoint>
<line-breakpoint enabled="true" suspend="THREAD" type="python-line">
<url>file://$PROJECT_DIR$/networks/seperating_adversarial_auto_encoder.py</url>
<line>23</line> <line>23</line>
<option name="timeStamp" value="27" /> <option name="timeStamp" value="52" />
</line-breakpoint>
<line-breakpoint enabled="true" suspend="THREAD" type="python-line">
<url>file://$PROJECT_DIR$/viz/viz_latent.py</url>
<line>67</line>
<option name="timeStamp" value="56" />
</line-breakpoint> </line-breakpoint>
</breakpoints> </breakpoints>
<default-breakpoints> <default-breakpoints>
<breakpoint type="python-exception"> <breakpoint type="python-exception">
<properties notifyOnTerminate="true" exception="BaseException"> <properties notifyOnlyOnFirst="true" notifyOnTerminate="true" ignoreLibraries="true" exception="BaseException">
<option name="ignoreLibraries" value="true" />
<option name="notifyOnTerminate" value="true" /> <option name="notifyOnTerminate" value="true" />
<option name="notifyOnlyOnFirst" value="true" />
</properties> </properties>
</breakpoint> </breakpoint>
</default-breakpoints> </default-breakpoints>
</breakpoint-manager> </breakpoint-manager>
</component> </component>
<component name="com.intellij.coverage.CoverageDataManagerImpl"> <component name="com.intellij.coverage.CoverageDataManagerImpl">
<SUITE FILE_PATH="coverage/ae_toolbox_torch$viz_latent.coverage" NAME="viz_latent Coverage Results" MODIFIED="1566541302103" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/viz" />
<SUITE FILE_PATH="coverage/ae_toolbox_torch$basic_ae_lightning_torch.coverage" NAME="basic_ae_lightning_torch Coverage Results" MODIFIED="1565937164457" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" /> <SUITE FILE_PATH="coverage/ae_toolbox_torch$basic_ae_lightning_torch.coverage" NAME="basic_ae_lightning_torch Coverage Results" MODIFIED="1565937164457" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
<SUITE FILE_PATH="coverage/ae_toolbox_torch$basic_ae_lightning.coverage" NAME="basic_ae_lightning Coverage Results" MODIFIED="1565956491159" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" /> <SUITE FILE_PATH="coverage/ae_toolbox_torch$basic_ae_lightning.coverage" NAME="basic_ae_lightning Coverage Results" MODIFIED="1565956491159" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
<SUITE FILE_PATH="coverage/ae_toolbox_torch$basic_vae_lightning.coverage" NAME="basic_vae_lightning Coverage Results" MODIFIED="1565955311009" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" /> <SUITE FILE_PATH="coverage/ae_toolbox_torch$basic_vae_lightning.coverage" NAME="basic_vae_lightning Coverage Results" MODIFIED="1565955311009" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
<SUITE FILE_PATH="coverage/ae_toolbox_torch$run_basic_ae.coverage" NAME="run_basic_ae Coverage Results" MODIFIED="1565966122607" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" /> <SUITE FILE_PATH="coverage/ae_toolbox_torch$run_basic_ae.coverage" NAME="run_basic_ae Coverage Results" MODIFIED="1565966122607" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
<SUITE FILE_PATH="coverage/ae_toolbox_torch$run_models.coverage" NAME="run_models Coverage Results" MODIFIED="1566306843739" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" /> <SUITE FILE_PATH="coverage/ae_toolbox_torch$run_models.coverage" NAME="run_models Coverage Results" MODIFIED="1566537126647" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$" />
<SUITE FILE_PATH="coverage/ae_toolbox_torch$dataset.coverage" NAME="dataset Coverage Results" MODIFIED="1565772669750" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/data" /> <SUITE FILE_PATH="coverage/ae_toolbox_torch$dataset.coverage" NAME="dataset Coverage Results" MODIFIED="1565772669750" SOURCE_PROVIDER="com.intellij.coverage.DefaultCoverageFileProvider" RUNNER="coverage.py" COVERAGE_BY_TEST_ENABLED="true" COVERAGE_TRACING_ENABLED="false" WORKING_DIRECTORY="$PROJECT_DIR$/data" />
</component> </component>
</project> </project>

3
dataset.py
View File

@ -163,7 +163,7 @@ class Trajectories(Dataset):
self.data = self.__init_data_(**kwargs) self.data = self.__init_data_(**kwargs)
pass pass
def __init_data_(self, **kwargs): def __init_data_(self, **kwargs: dict):
dataDict = dict() dataDict = dict()
for key, val in kwargs.items(): for key, val in kwargs.items():
if key in self.isovistMeasures: if key in self.isovistMeasures:
@ -177,6 +177,7 @@ class Trajectories(Dataset):
return data return data
def __iter__(self): def __iter__(self):
# FixMe: is that correct?
for i in range(len(self)): for i in range(len(self)):
yield self[i] yield self[i]

4
networks/adverserial_auto_encoder.py
View File

@ -10,7 +10,7 @@ class AdversarialAutoEncoder(AutoEncoder):
def __init__(self, *args, **kwargs): def __init__(self, *args, **kwargs):
super(AdversarialAutoEncoder, self).__init__(*args, **kwargs) super(AdversarialAutoEncoder, self).__init__(*args, **kwargs)
self.discriminator = Discriminator(self.latent_dim, self.dataParams) self.discriminator = Discriminator(self.latent_dim, self.features)
def forward(self, batch): def forward(self, batch):
# Encoder # Encoder
@ -18,7 +18,7 @@ class AdversarialAutoEncoder(AutoEncoder):
z = self.encoder(batch) z = self.encoder(batch)
# Decoder # Decoder
# First repeat the data accordingly to the batch size # First repeat the data accordingly to the batch size
z_repeatet = Repeater((batch.shape[0], self.dataParams['size'], -1))(z) z_repeatet = Repeater((batch.shape[0], batch.shape[1], -1))(z)
x_hat = self.decoder(z_repeatet) x_hat = self.decoder(z_repeatet)
return z, x_hat return z, x_hat

9
networks/auto_encoder.py
View File

@ -7,12 +7,13 @@ from torch import Tensor
# Basic AE-Implementation # Basic AE-Implementation
class AutoEncoder(AbstractNeuralNetwork, ABC): class AutoEncoder(AbstractNeuralNetwork, ABC):
def __init__(self, latent_dim: int, dataParams: dict, **kwargs): def __init__(self, latent_dim: int=0, features: int = 0, **kwargs):
assert latent_dim and features
super(AutoEncoder, self).__init__() super(AutoEncoder, self).__init__()
self.dataParams = dataParams
self.latent_dim = latent_dim self.latent_dim = latent_dim
self.features = features
self.encoder = Encoder(self.latent_dim) self.encoder = Encoder(self.latent_dim)
self.decoder = Decoder(self.latent_dim, self.dataParams['features']) self.decoder = Decoder(self.latent_dim, self.features)
def forward(self, batch: Tensor): def forward(self, batch: Tensor):
# Encoder # Encoder
@ -20,7 +21,7 @@ class AutoEncoder(AbstractNeuralNetwork, ABC):
z = self.encoder(batch) z = self.encoder(batch)
# Decoder # Decoder
# First repeat the data accordingly to the batch size # First repeat the data accordingly to the batch size
z_repeatet = Repeater((batch.shape[0], self.dataParams['size'], -1))(z) z_repeatet = Repeater((batch.shape[0], batch.shape[1], -1))(z)
x_hat = self.decoder(z_repeatet) x_hat = self.decoder(z_repeatet)
return z, x_hat return z, x_hat

10
networks/modules.py
View File

@ -131,13 +131,13 @@ class AvgDimPool(Module):
# Generators, Decoders, Encoders, Discriminators # Generators, Decoders, Encoders, Discriminators
class Discriminator(Module): class Discriminator(Module):
def __init__(self, latent_dim, dataParams, dropout=.0, activation=ReLU): def __init__(self, latent_dim, features, dropout=.0, activation=ReLU):
super(Discriminator, self).__init__() super(Discriminator, self).__init__()
self.dataParams = dataParams self.features = features
self.latent_dim = latent_dim self.latent_dim = latent_dim
self.l1 = Linear(self.latent_dim, self.dataParams['features'] * 10) self.l1 = Linear(self.latent_dim, self.features * 10)
self.l2 = Linear(self.dataParams['features'] * 10, self.dataParams['features'] * 20) self.l2 = Linear(self.features * 10, self.features * 20)
self.lout = Linear(self.dataParams['features'] * 20, 1) self.lout = Linear(self.features * 20, 1)
self.dropout = Dropout(dropout) self.dropout = Dropout(dropout)
self.activation = activation() self.activation = activation()
self.sigmoid = Sigmoid() self.sigmoid = Sigmoid()

12
networks/seperating_adversarial_auto_encoder.py
View File

@ -6,17 +6,17 @@ import torch
class SeperatingAdversarialAutoEncoder(Module): class SeperatingAdversarialAutoEncoder(Module):
def __init__(self, latent_dim, dataParams, **kwargs): def __init__(self, latent_dim, features, **kwargs):
assert latent_dim % 2 == 0, f'Your latent space needs to be even, not odd, but was: "{latent_dim}"' assert latent_dim % 2 == 0, f'Your latent space needs to be even, not odd, but was: "{latent_dim}"'
super(SeperatingAdversarialAutoEncoder, self).__init__() super(SeperatingAdversarialAutoEncoder, self).__init__()
self.latent_dim = latent_dim self.latent_dim = latent_dim
self.dataParams = dataParams self.features = features
self.spatial_encoder = PoolingEncoder(self.latent_dim // 2) self.spatial_encoder = PoolingEncoder(self.latent_dim // 2)
self.temporal_encoder = Encoder(self.latent_dim // 2) self.temporal_encoder = Encoder(self.latent_dim // 2)
self.decoder = Decoder(self.latent_dim, self.dataParams['features']) self.decoder = Decoder(self.latent_dim, self.features)
self.spatial_discriminator = Discriminator(self.latent_dim // 2, self.dataParams) self.spatial_discriminator = Discriminator(self.latent_dim // 2, self.features)
self.temporal_discriminator = Discriminator(self.latent_dim // 2, self.dataParams) self.temporal_discriminator = Discriminator(self.latent_dim // 2, self.features)
def forward(self, batch): def forward(self, batch):
# Encoder # Encoder
@ -25,7 +25,7 @@ class SeperatingAdversarialAutoEncoder(Module):
# Decoder # Decoder
# First repeat the data accordingly to the batch size # First repeat the data accordingly to the batch size
z_concat = torch.cat((z_spatial, z_temporal), dim=-1) z_concat = torch.cat((z_spatial, z_temporal), dim=-1)
z_repeatet = Repeater((batch.shape[0], self.dataParams['size'], -1))(z_concat) z_repeatet = Repeater((batch.shape[0], batch.shape[1], -1))(z_concat)
x_hat = self.decoder(z_repeatet) x_hat = self.decoder(z_repeatet)
return z_spatial, z_temporal, x_hat return z_spatial, z_temporal, x_hat

11
networks/variational_auto_encoder.py
View File

@ -10,12 +10,13 @@ class VariationalAutoEncoder(AbstractNeuralNetwork, ABC):
def name(self): def name(self):
return self.__class__.__name__ return self.__class__.__name__
def __init__(self, dataParams, **kwargs): def __init__(self, latent_dim=0, features=0, **kwargs):
assert latent_dim and features
super(VariationalAutoEncoder, self).__init__() super(VariationalAutoEncoder, self).__init__()
self.dataParams = dataParams self.features = features
self.latent_dim = kwargs.get('latent_dim', 2) self.latent_dim = latent_dim
self.encoder = Encoder(self.latent_dim, variational=True) self.encoder = Encoder(self.latent_dim, variational=True)
self.decoder = Decoder(self.latent_dim, self.dataParams['features'], variational=True) self.decoder = Decoder(self.latent_dim, self.features, variational=True)
@staticmethod @staticmethod
def reparameterize(mu, logvar): def reparameterize(mu, logvar):
@ -27,7 +28,7 @@ class VariationalAutoEncoder(AbstractNeuralNetwork, ABC):
def forward(self, batch): def forward(self, batch):
mu, logvar = self.encoder(batch) mu, logvar = self.encoder(batch)
z = self.reparameterize(mu, logvar) z = self.reparameterize(mu, logvar)
repeat = Repeater((batch.shape[0], self.dataParams['size'], -1)) repeat = Repeater((batch.shape[0], batch.shape[1], -1))
x_hat = self.decoder(repeat(z)) x_hat = self.decoder(repeat(z))
return x_hat, mu, logvar return x_hat, mu, logvar

66
run_models.py
View File

@ -14,21 +14,35 @@ from torch.nn import BatchNorm1d
from pytorch_lightning import Trainer from pytorch_lightning import Trainer
from test_tube import Experiment from test_tube import Experiment
from argparse import Namespace
from argparse import ArgumentParser
args = ArgumentParser()
args.add_argument('step')
args.add_argument('features')
args.add_argument('size')
args.add_argument('latent_dim')
# ToDo: How to implement this better? # ToDo: How to implement this better?
# other_classes = [AutoEncoder, AutoEncoderLightningOverrides] # other_classes = [AutoEncoder, AutoEncoderLightningOverrides]
class Model(VariationalAutoEncoderLightningOverrides, LightningModule): class Model(AutoEncoderLightningOverrides, LightningModule):
def __init__(self, dataParams: dict): def __init__(self, latent_dim=0, size=0, step=0, features=0, **kwargs):
assert all([x in args for x in ['step', 'size', 'latent_dim', 'features']])
self.size = args.size
self.latent_dim = args.latent_dim
self.features = args.features
self.step = args.step
super(Model, self).__init__() super(Model, self).__init__()
self.dataParams = dataParams self.network = AutoEncoder(self.latent_dim, self.features)
self.network = VariationalAutoEncoder(self.dataParams)
def configure_optimizers(self): def configure_optimizers(self):
return [Adam(self.parameters(), lr=0.02)] return [Adam(self.parameters(), lr=0.02)]
@data_loader @data_loader
def tng_dataloader(self): def tng_dataloader(self):
return DataLoader(DataContainer('data', **self.dataParams), shuffle=True, batch_size=100) return DataLoader(DataContainer('data', self.size, self.step), shuffle=True, batch_size=100)
class AdversarialModel(AdversarialAELightningOverrides, LightningModule): class AdversarialModel(AdversarialAELightningOverrides, LightningModule):
@ -37,11 +51,15 @@ class AdversarialModel(AdversarialAELightningOverrides, LightningModule):
def name(self): def name(self):
return self.network.name return self.network.name
def __init__(self, dataParams: dict): def __init__(self, args: Namespace, **kwargs):
assert all([x in args for x in ['step', 'size', 'latent_dim', 'features']])
self.size = args.size
self.latent_dim = args.latent_dim
self.features = args.features
self.step = args.step
super(AdversarialModel, self).__init__() super(AdversarialModel, self).__init__()
self.dataParams = dataParams
self.normal = Normal(0, 1) self.normal = Normal(0, 1)
self.network = AdversarialAutoEncoder(self.dataParams) self.network = AdversarialAutoEncoder(self.latent_dim, self.features)
pass pass
# This is Fucked up, why do i need to put an additional empty list here? # This is Fucked up, why do i need to put an additional empty list here?
@ -52,17 +70,20 @@ class AdversarialModel(AdversarialAELightningOverrides, LightningModule):
@data_loader @data_loader
def tng_dataloader(self): def tng_dataloader(self):
return DataLoader(DataContainer('data', **self.dataParams), shuffle=True, batch_size=100) return DataLoader(DataContainer('data', self.size, self.step), shuffle=True, batch_size=100)
class SeparatingAdversarialModel(SeparatingAdversarialAELightningOverrides, LightningModule): class SeparatingAdversarialModel(SeparatingAdversarialAELightningOverrides, LightningModule):
def __init__(self, latent_dim, dataParams: dict): def __init__(self, args: Namespace, **kwargs):
assert all([x in args for x in ['step', 'size', 'latent_dim', 'features']])
self.size = args.size
self.latent_dim = args.latent_dim
self.features = args.features
self.step = args.step
super(SeparatingAdversarialModel, self).__init__() super(SeparatingAdversarialModel, self).__init__()
self.latent_dim = latent_dim
self.dataParams = dataParams
self.normal = Normal(0, 1) self.normal = Normal(0, 1)
self.network = SeperatingAdversarialAutoEncoder(self.latent_dim, self.dataParams) self.network = SeperatingAdversarialAutoEncoder(self.latent_dim, self.features, **kwargs)
pass pass
# This is Fucked up, why do i need to put an additional empty list here? # This is Fucked up, why do i need to put an additional empty list here?
@ -78,22 +99,24 @@ class SeparatingAdversarialModel(SeparatingAdversarialAELightningOverrides, Ligh
@data_loader @data_loader
def tng_dataloader(self): def tng_dataloader(self):
return DataLoader(DataContainer('data', **self.dataParams), shuffle=True, batch_size=100) num_workers = os.cpu_count() // 2
return DataLoader(DataContainer('data', self.size, self.step), shuffle=True, batch_size=100, num_workers=num_workers)
if __name__ == '__main__': if __name__ == '__main__':
features = 6 features = 6
latent_dim = 4 tag_dict = dict(features=features, latent_dim=4, size=5, step=6, refresh=False,
model = SeparatingAdversarialModel(latent_dim=latent_dim, dataParams=dict(refresh=False, size=5, step=5, transforms=[BatchNorm1d(features)])
features=features, transforms=[BatchNorm1d(features)] arguments = args.parse_args()
) arguments.__dict__.update(tag_dict)
)
model = SeparatingAdversarialModel(arguments)
# PyTorch summarywriter with a few bells and whistles # PyTorch summarywriter with a few bells and whistles
outpath = os.path.join(os.getcwd(), 'output', model.name, time.asctime().replace(' ', '_').replace(':', '-')) outpath = os.path.join(os.getcwd(), 'output', model.name, time.asctime().replace(' ', '_').replace(':', '-'))
os.makedirs(outpath, exist_ok=True) os.makedirs(outpath, exist_ok=True)
exp = Experiment(save_dir=outpath) exp = Experiment(save_dir=outpath)
exp.tag(tag_dict=tag_dict)
from pytorch_lightning.callbacks import ModelCheckpoint from pytorch_lightning.callbacks import ModelCheckpoint
@ -101,9 +124,8 @@ if __name__ == '__main__':
filepath=os.path.join(outpath, 'weights.ckpt'), filepath=os.path.join(outpath, 'weights.ckpt'),
save_best_only=True, save_best_only=True,
verbose=True, verbose=True,
monitor='val_loss', monitor='tng_loss', # val_loss
mode='min', mode='min',
) )
trainer = Trainer(experiment=exp, checkpoint_callback=checkpoint_callback, max_nb_epochs=15) # gpus=[0...LoL] trainer = Trainer(experiment=exp, checkpoint_callback=checkpoint_callback, max_nb_epochs=15) # gpus=[0...LoL]

53
viz/viz_latent.py
View File

@ -4,6 +4,8 @@ import torch
from torch.utils.data import DataLoader from torch.utils.data import DataLoader
from pytorch_lightning import data_loader from pytorch_lightning import data_loader
from dataset import DataContainer from dataset import DataContainer
from collections import defaultdict
from tqdm import tqdm
import os import os
from sklearn.manifold import TSNE from sklearn.manifold import TSNE
@ -12,30 +14,28 @@ from sklearn.decomposition import PCA
import seaborn as sns; sns.set() import seaborn as sns; sns.set()
import matplotlib.pyplot as plt import matplotlib.pyplot as plt
from run_models import SeparatingAdversarialModel from run_models import *
path = 'output'
mylightningmodule = 'weired name, loaded from disk'
# FIXME: How to store hyperparamters in testtube element?
def search_for_weights(folder): def search_for_weights(folder):
while not os.path.exists(folder):
if len(os.path.split(folder)) >= 50:
raise FileNotFoundError(f'The folder "{folder}" could not be found')
folder = os.path.join(os.pardir, folder)
for element in os.scandir(folder): for element in os.scandir(folder):
if os.path.exists(element): if os.path.exists(element):
if element.is_dir(): if element.is_dir():
search_for_weights(element.path) search_for_weights(element.path)
elif element.is_file() and element.name.endswith('.ckpt'): elif element.is_file() and element.name.endswith('.ckpt'):
load_and_viz(element) load_and_predict(element)
else: else:
continue continue
def load_and_viz(path_like_element): def load_and_predict(path_like_element):
# Define Loop to search for models and folder with visualizations # Define Loop to search for models and folder with visualizations
pretrained_model = SeparatingAdversarialModel.load_from_metrics( model = globals()[path_like_element.path.split(os.sep)[-3]]
pretrained_model = model.load_from_metrics(
weights_path=path_like_element.path, weights_path=path_like_element.path,
tags_csv=os.path.join(os.path.dirname(path_like_element), 'default', 'version_0', 'meta_tags.csv'), tags_csv=os.path.join(os.path.dirname(path_like_element), 'default', 'version_0', 'meta_tags.csv'),
on_gpu=True if torch.cuda.is_available() else False, on_gpu=True if torch.cuda.is_available() else False,
@ -46,19 +46,26 @@ def load_and_viz(path_like_element):
pretrained_model.eval() pretrained_model.eval()
pretrained_model.freeze() pretrained_model.freeze()
# Load the data fpr prediction # Load the data for prediction
dataset = DataContainer('data', 5, 5) dataset = DataContainer(os.path.join(os.pardir, 'data'), 5, 5)
# Do the inference # Do the inference
predictions = [] prediction_dict = defaultdict(list)
for i in range(len(dataset)): for i in tqdm(range(len(dataset)), total=len(dataset)):
z, _ = pretrained_model(dataset[i]) p_X = pretrained_model(dataset[i].unsqueeze(0))
predictions.append(z) for idx in range(len(p_X) - 1):
predictions = torch.cat(predictions) prediction_dict[idx].append(p_X[idx])
if predictions.shape[-1] <= 1:
predictions = [torch.cat(prediction).detach().numpy() for prediction in prediction_dict.values()]
for prediction in predictions:
viz_latent(prediction)
def viz_latent(prediction):
if prediction.shape[-1] <= 1:
raise ValueError('How did this happen?') raise ValueError('How did this happen?')
elif predictions.shape[-1] == 2: elif prediction.shape[-1] == 2:
ax = sns.scatterplot(x=predictions[:, 0], y=predictions[:, 1]) ax = sns.scatterplot(x=prediction[:, 0], y=prediction[:, 1])
plt.show() plt.show()
return ax return ax
else: else:
@ -69,3 +76,7 @@ def load_and_viz(path_like_element):
tsne_plot = sns.scatterplot(x=predictions_tsne[:, 0], y=predictions_tsne[:, 1], ax=axs[1]) tsne_plot = sns.scatterplot(x=predictions_tsne[:, 0], y=predictions_tsne[:, 1], ax=axs[1])
plt.show() plt.show()
return fig, axs, pca_plot, tsne_plot return fig, axs, pca_plot, tsne_plot
if __name__ == '__main__':
path = 'output'
search_for_weights(path)