Merge branch 'master' into onnx-dmytro

pom.xml

@@ -245,3 +245,4 @@
         </snapshotRepository>
     </distributionManagement>
 </project>
+

src/main/java/de/monticore/lang/monticar/cnnarch/generator/ArchitectureElementData.java

@@ -90,6 +90,10 @@ public class ArchitectureElementData {
         return getLayerSymbol().getIntTupleValue(AllPredefinedLayers.STRIDE_NAME).get();
     }
 
+    public String getPdf() { return getLayerSymbol().getStringValue(AllPredefinedLayers.PDF_NAME).get(); }
+
+    public int getNumEmbeddings() { return getLayerSymbol().getIntValue(AllPredefinedLayers.NUM_EMBEDDINGS_NAME).get();
+}
     public int getGroups(){
         return getLayerSymbol().getIntValue(AllPredefinedLayers.GROUPS_NAME).get();
     }
@@ -314,7 +318,11 @@ public class ArchitectureElementData {
     public int getValuesDim(){
         return getLayerSymbol().getIntValue(AllPredefinedLayers.VALUES_DIM_NAME).get();
     }
-    
+
+    public int getNodes(){
+        return getLayerSymbol().getIntValue(AllPredefinedLayers.NODES_NAME).get();
+    }
+
     @Nullable
     public List<Integer> getPadding(){
 

src/main/java/de/monticore/lang/monticar/cnnarch/generator/CNNArchTemplateController.java

@@ -106,6 +106,7 @@ public abstract class CNNArchTemplateController {
     public boolean containsAdaNet(){
         return this.architecture.containsAdaNet();
     }
+
     public String getName(ArchitectureElementSymbol layer){
         return nameManager.getName(layer);
     }

src/main/java/de/monticore/lang/monticar/cnnarch/generator/ConfigurationData.java

@@ -6,6 +6,7 @@ import com.google.common.collect.Maps;
 import de.monticore.lang.monticar.cnnarch.generator.annotations.ArchitectureAdapter;
 import de.monticore.lang.monticar.cnnarch.generator.annotations.Range;
 import de.monticore.lang.monticar.cnnarch.generator.training.RlAlgorithm;
+import de.monticore.lang.monticar.cnnarch.generator.training.NetworkType;
 import de.monticore.lang.monticar.cnnarch.generator.training.TrainingComponentsContainer;
 import de.monticore.lang.monticar.cnnarch.generator.training.TrainingConfiguration;
 
@@ -39,6 +40,10 @@ public abstract class ConfigurationData {
         return trainingConfiguration.isGanLearning();
     }
 
+    public Boolean isVaeLearning() {
+        return trainingConfiguration.isVaeLearning();
+    }
+
     public Boolean isReinforcementLearning() {
         return trainingConfiguration.isReinforcementLearning();
     }
@@ -73,6 +78,16 @@ public abstract class ConfigurationData {
         return loadPretrainedOpt.orElse(null);
     }
 
+    public Double getKlLossWeight() {
+        Optional<Double> klLossWeightOpt = trainingConfiguration.getKlLossWeight();
+        return klLossWeightOpt.orElse(null);
+    }
+
+    public String getReconLossName() {
+        Optional<String> reconLossNameOpt = trainingConfiguration.getReconLossName();
+        return reconLossNameOpt.orElse(null);
+    }
+
     // COMPARE WITH CNNTRAIN IMPLEMENTATION IN GluonConfigurationData
     public Boolean getPreprocessor() {
         Optional<String> preprocessorOpt = trainingConfiguration.getPreprocessor();
@@ -89,6 +104,21 @@ public abstract class ConfigurationData {
         return onnxExport.orElse(null);
     }
 
+    public Boolean getMultiGraph() {
+        Optional<Boolean> multiGraphOpt = trainingConfiguration.getMultiGraph();
+        return multiGraphOpt.orElse(null);
+    }
+
+    public List<Integer> getTrainMask() {
+        Optional<List<Integer>> trainMaskOpt = trainingConfiguration.getTrainMask();
+        return trainMaskOpt.orElse(null);
+    }
+
+    public List<Integer> getTestMask() {
+        Optional<List<Integer>> testMaskOpt = trainingConfiguration.getTestMask();
+        return testMaskOpt.orElse(null);
+    }
+
     public Boolean getShuffleData() {
         Optional<Boolean> shuffleDataOpt = trainingConfiguration.getShuffleData();
         return shuffleDataOpt.orElse(null);
@@ -352,7 +382,11 @@ public abstract class ConfigurationData {
 //    public Map<String, Map<String, Object>> getConstraintLosses() { // TODO
 //        return getMultiParamMapEntry(CONSTRAINT_LOSS, "name");
 //    }
-
+    public String getSelfPlay() { // added Parameter self_play for cooperative driving
+        Optional<String> selfPlay = trainingConfiguration.getSelfPlay();
+        return selfPlay.orElse(null);
+    }
+    
     public String getRlAlgorithm() {
         Optional<RlAlgorithm> rlAlgorithmOpt = trainingConfiguration.getRlAlgorithm();
         if (!rlAlgorithmOpt.isPresent()) {
@@ -370,6 +404,16 @@ public abstract class ConfigurationData {
         return DQN;
     }
 
+    public String getNetworkType() {
+        Optional<NetworkType> networkTypeOpt = trainingConfiguration.getNetworkType();
+
+        NetworkType networkType = networkTypeOpt.get();
+        if (networkType.equals(NetworkType.GNN)) {
+            return GNN;
+        }
+        return null;
+    }
+
 //    protected Object getDefaultValueOrElse(String parameterKey, Object elseValue) {
 //        if (schema == null) {
 //            return elseValue;
@@ -631,4 +675,4 @@ public abstract class ConfigurationData {
         }
         return object.toString();
     }
-}
\ No newline at end of file
+}

src/main/java/de/monticore/lang/monticar/cnnarch/generator/training/LearningMethod.java

@@ -4,7 +4,8 @@ public enum LearningMethod {
 
     SUPERVISED("supervised"),
     REINFORCEMENT("reinforcement"),
-    GAN("gan");
+    GAN("gan"),
+    VAE("vae");
 
     String method;
 

src/main/java/de/monticore/lang/monticar/cnnarch/generator/training/NetworkType.java

+package de.monticore.lang.monticar.cnnarch.generator.training;
+
+public enum NetworkType {
+
+    GNN("gnn");
+
+    String type;
+
+    NetworkType(String type) {
+        this.type = type;
+    }
+
+    public static NetworkType networkType(String type) {
+        for (NetworkType nt : values()) {
+            if (nt.type.equals(type)) {
+                return nt;
+            }
+        }
+        throw new IllegalArgumentException(String.valueOf(type));
+    }
+
+    public String getType() {
+        return type;
+    }
+}
\ No newline at end of file

src/main/java/de/monticore/lang/monticar/cnnarch/generator/training/TrainingComponentsContainer.java

@@ -20,6 +20,8 @@ public class TrainingComponentsContainer {
     private ArchitectureAdapter actorNetwork;
     private ArchitectureAdapter criticNetwork;
     private ArchitectureAdapter generatorNetwork;
+    private ArchitectureAdapter encoderNetwork;
+    private ArchitectureAdapter decoderNetwork;
     private ArchitectureAdapter discriminatorNetwork;
     private ArchitectureAdapter qNetwork;
     private EMAComponentInstanceSymbol rewardFunction;
@@ -43,11 +45,18 @@ public class TrainingComponentsContainer {
         return Optional.ofNullable(generatorNetwork);
     }
 
-
     public Optional<ArchitectureAdapter> getDiscriminatorNetwork() {
         return Optional.ofNullable(discriminatorNetwork);
     }
 
+    public Optional<ArchitectureAdapter> getDecoderNetwork() {
+        return Optional.ofNullable(decoderNetwork);
+    }
+
+    public Optional<ArchitectureAdapter> getEncoderNetwork() {
+        return Optional.ofNullable(encoderNetwork);
+    }
+
     public Optional<ArchitectureAdapter> getQNetwork() {
         return Optional.ofNullable(qNetwork);
     }
@@ -80,6 +89,8 @@ public class TrainingComponentsContainer {
             }
         } else if (trainingConfiguration.isGanLearning()) {
             setGeneratorNetwork(trainedArchitecture);
+        } else if (trainingConfiguration.isVaeLearning()) {
+            setDecoderNetwork(trainedArchitecture);
         }
     }
 
@@ -108,6 +119,16 @@ public class TrainingComponentsContainer {
         addTrainingComponent(QNETWORK, qNetwork);
     }
 
+    public void setDecoderNetwork(ArchitectureAdapter decoderNetwork) {
+        this.decoderNetwork = decoderNetwork;
+        addTrainingComponent(DECODER, decoderNetwork);
+    }
+
+    public void setEncoderNetwork(ArchitectureAdapter encoderNetwork) {
+        this.encoderNetwork = encoderNetwork;
+        addTrainingComponent(ENCODER, encoderNetwork);
+    }
+
     public void setRewardFunction(EMAComponentInstanceSymbol rewardFunction) {
         this.rewardFunction = rewardFunction;
         addTrainingComponent(REWARD_FUNCTION, rewardFunction.getComponentType());

src/main/java/de/monticore/lang/monticar/cnnarch/generator/training/TrainingConfiguration.java

@@ -7,10 +7,7 @@ import conflang._symboltable.ConfigurationEntrySymbol;
 import conflang._symboltable.NestedConfigurationEntrySymbol;
 import schemalang._symboltable.SchemaDefinitionSymbol;
 
-import java.util.Collection;
-import java.util.List;
-import java.util.Map;
-import java.util.Optional;
+import java.util.*;
 
 import static de.monticore.lang.monticar.cnnarch.generator.training.TrainingParameterConstants.*;
 
@@ -50,6 +47,10 @@ public class TrainingConfiguration {
         return getParameterValue(CONTEXT);
     }
 
+    public Optional<String> getSelfPlay() {
+        return getParameterValue(SELF_PLAY);
+    }
+
     public Optional<LearningMethod> getLearningMethod() {
         Optional<ConfigurationEntry> learningMethodOpt =
                 configurationSymbol.getConfigurationEntry(LEARNING_METHOD);
@@ -83,6 +84,15 @@ public class TrainingConfiguration {
         return LearningMethod.GAN.equals(learningMethod);
     }
 
+    public boolean isVaeLearning() {
+        Optional<LearningMethod> learningMethodOpt = getLearningMethod();
+        if (!learningMethodOpt.isPresent()) {
+            return false; // Not correct here to return false..
+        }
+        LearningMethod learningMethod = learningMethodOpt.get();
+        return LearningMethod.VAE.equals(learningMethod);
+    }
+
     public boolean isReinforcementLearning() {
         Optional<LearningMethod> learningMethodOpt = getLearningMethod();
         if (!learningMethodOpt.isPresent()) {
@@ -101,6 +111,15 @@ public class TrainingConfiguration {
         return Optional.of(RlAlgorithm.rlAlgorithm(rlAlgorithm));
     }
 
+    public Optional<NetworkType> getNetworkType() {
+        Optional<ConfigurationEntry> networkTypeOpt = configurationSymbol.getConfigurationEntry(NETWORK_TYPE);
+        if (!networkTypeOpt.isPresent()) {
+            return Optional.empty();
+        }
+        String networkType = (String) networkTypeOpt.get().getValue();
+        return Optional.of(NetworkType.networkType(networkType));
+    }
+
     public Optional<Integer> getBatchSize() {
         return getParameterValue(BATCH_SIZE);
     }
@@ -141,6 +160,18 @@ public class TrainingConfiguration {
         return getParameterValue(SHUFFLE_DATA);
     }
 
+    public Optional<Boolean> getMultiGraph() {
+        return getParameterValue(MULTI_GRAPH);
+    }
+
+    public Optional<List<Integer>> getTrainMask() {
+        return getParameterValue(TRAIN_MASK);
+    }
+
+    public Optional<List<Integer>> getTestMask() {
+        return getParameterValue(TEST_MASK);
+    }
+
     public Optional<Double> getClipGlobalGradNorm() {
         return getParameterValue(CLIP_GLOBAL_GRAD_NORM);
     }
@@ -353,6 +384,20 @@ public class TrainingConfiguration {
         return getObjectParameterParameters(DISCRIMINATOR_OPTIMIZER);
     }
 
+    public Boolean hasEncoderName() {
+        return hasParameter(ENCODER);
+    }
+
+    public Optional<String> getEncoderName() {
+        return getObjectParameterValue(ENCODER);
+    }
+
+    public Optional<String> getReconLossName() {
+        return getObjectParameterValue(RECON_LOSS);
+    }
+
+    public Optional<Double> getKlLossWeight() { return getParameterValue(KL_LOSS_WEIGHT); }
+
     public boolean hasStrategy() {
         return hasParameter(STRATEGY);
     }
@@ -513,4 +558,5 @@ public class TrainingConfiguration {
         }
         return keyValues;
     }
+
 }
\ No newline at end of file

src/main/java/de/monticore/lang/monticar/cnnarch/generator/training/TrainingParameterConstants.java

@@ -15,6 +15,7 @@ public class TrainingParameterConstants {
     public static final String SUPERVISED = "supervised";
     public static final String REINFORCEMENT = "reinforcement";
     public static final String GAN = "gan";
+    public static final String VAE = "vae";
 
     /*
      * Optimizers
@@ -41,6 +42,7 @@ public class TrainingParameterConstants {
     public static final String NORMALIZE = "normalize";
     public static final String CONTEXT = "context";
     public static final String SHUFFLE_DATA = "shuffle_data";
+
     public static final String CLIP_GLOBAL_GRAD_NORM = "clip_global_grad_norm";
     public static final String USE_TEACHER_FORCING = "use_teacher_forcing";
     public static final String SAVE_ATTENTION_IMAGE = "save_attention_image";
@@ -56,6 +58,13 @@ public class TrainingParameterConstants {
     public static final String DQN = "dqn";
     public static final String DDPG = "ddpg";
     public static final String TD3 = "td3";
+    public static final String SELF_PLAY = "self_play";
+
+    public static final String MULTI_GRAPH = "multi_graph";
+    public static final String TRAIN_MASK = "train_mask";
+    public static final String TEST_MASK = "test_mask";
+    public static final String GNN = "gnn";
+    public static final String NETWORK_TYPE = "network_type";
 
     public static final String LEARNING_METHOD = "learning_method";
     public static final String EVAL_METRIC = "eval_metric";
@@ -111,4 +120,9 @@ public class TrainingParameterConstants {
     public static final String GENERATOR_LOSS_WEIGHT = "generator_loss_weight";
     public static final String DISCRIMINATOR_LOSS_WEIGHT = "discriminator_loss_weight";
     public static final String PRINT_IMAGES = "print_images";
+
+    public static final String ENCODER = "encoder";
+    public static final String DECODER = "decoder";
+    public static final String KL_LOSS_WEIGHT = "kl_loss_weight";
+    public static final String RECON_LOSS = "reconstruction_loss";
 }
\ No newline at end of file

src/main/resources/schemas/GNN.scm

+/* (c) https://github.com/MontiCore/monticore */
+
+schema GNN extends Supervised {
+    train_mask: Z*
+    test_mask: Z*
+    multi_graph: B
+}

src/main/resources/schemas/General.scm

@@ -4,7 +4,7 @@ import Optimizer;
 schema General {
 
     learning_method = supervised: schema {
-        supervised, reinforcement, gan;
+        supervised, reinforcement, gan, vae;
     }
 
     context: enum {

src/main/resources/schemas/Reinforcement.scm

@@ -8,6 +8,10 @@ schema Reinforcement extends General {
         dqn, ddpg, td3;
     }
 
+    self_play: enum {
+	no, yes;
+    }
+
     agent_name: string
     num_episodes = 50: N1
     num_max_steps = 99999: N
@@ -21,4 +25,4 @@ schema Reinforcement extends General {
     actor_optimizer: optimizer_type
     environment: environment_type!
     replay_memory = buffer: replay_memory_type
-}
\ No newline at end of file
+}

src/main/resources/schemas/Supervised.scm

@@ -4,6 +4,10 @@ import Loss;
 
 schema Supervised extends General {
 
+    network_type: schema {
+        gnn;
+    }
+
     batch_size: N1
     num_epoch: N
     normalize: B

src/main/resources/schemas/VAE.scm

+/* (c) https://github.com/MontiCore/monticore */
+import Optimizer;
+
+schema VAE extends General {
+
+  reference-model: referencemodels.vae.VAE, referencemodels.vae.CVAE
+  batch_size: N1
+  num_epoch: N1
+  normalize: B
+  checkpoint_period = 5: N
+  load_checkpoint: B
+  load_pretrained: B
+  log_period: N
+  reconstruction_loss = mse: reconLoss_type
+  print_images = false: B
+  kl_loss_weight: Q
+
+  reconLoss_type {
+     values:
+         bce,
+         mse;
+   }
+
+}
\ No newline at end of file

src/main/resources/schemas/referencemodels/vae/CVAE.ema

+/* (c) https://github.com/MontiCore/monticore */
+package referencemodels.vae;
+
+component CVAE {
+
+    component Encoder {
+        ports
+            in X data,
+            in W^{1} label,
+            out D encoding;
+    }
+
+    component Decoder {
+        ports
+            in D encoding,
+            in W^{1} label,
+            out X data;
+    }
+
+    instance Encoder encoder;
+    instance Decoder decoder;
+
+    connect encoder.encoding -> decoder.encoding;
+}
\ No newline at end of file

src/main/resources/schemas/referencemodels/vae/VAE.ema

+/* (c) https://github.com/MontiCore/monticore */
+package referencemodels.vae;
+
+component VAE {
+
+    component Encoder{
+        ports
+            in X data,
+            out D encoding;
+    }
+
+    component Decoder{
+        ports
+            in D encoding,
+            out X data;
+    }
+
+    instance Encoder encoder;
+    instance Decoder decoder;
+
+    connect encoder.encoding -> decoder.encoding;
+}
\ No newline at end of file