- calcGradient(IActorCritic, Stack<MiniTrans<Integer>>) - Method in class org.deeplearning4j.rl4j.learning.async.a3c.discrete.A3CThreadDiscrete
-
calc the gradients based on the n-step rewards
- calcGradient(NN, Stack<MiniTrans<Integer>>) - Method in class org.deeplearning4j.rl4j.learning.async.AsyncThreadDiscrete
-
- calcGradient(IDQN, Stack<MiniTrans<Integer>>) - Method in class org.deeplearning4j.rl4j.learning.async.nstep.discrete.AsyncNStepQLearningThreadDiscrete
-
- CartpoleNative - Class in org.deeplearning4j.rl4j.mdp
-
- CartpoleNative() - Constructor for class org.deeplearning4j.rl4j.mdp.CartpoleNative
-
- CartpoleNative(int) - Constructor for class org.deeplearning4j.rl4j.mdp.CartpoleNative
-
- CartpoleNative.KinematicsIntegrators - Enum in org.deeplearning4j.rl4j.mdp
-
- CartpoleNative.State - Class in org.deeplearning4j.rl4j.mdp
-
- cg - Variable in class org.deeplearning4j.rl4j.network.ac.ActorCriticCompGraph
-
- ChannelStackPoolContentAssembler - Class in org.deeplearning4j.rl4j.observation.preprocessor.pooling
-
ChannelStackPoolContentAssembler is used with the PoolingDataSetPreProcessor.
- ChannelStackPoolContentAssembler() - Constructor for class org.deeplearning4j.rl4j.observation.preprocessor.pooling.ChannelStackPoolContentAssembler
-
- CircularFifoObservationPool - Class in org.deeplearning4j.rl4j.observation.preprocessor.pooling
-
CircularFifoObservationPool is used with the PoolingDataSetPreProcessor.
- CircularFifoObservationPool() - Constructor for class org.deeplearning4j.rl4j.observation.preprocessor.pooling.CircularFifoObservationPool
-
- CircularFifoObservationPool(int) - Constructor for class org.deeplearning4j.rl4j.observation.preprocessor.pooling.CircularFifoObservationPool
-
- CircularFifoObservationPool.Builder - Class in org.deeplearning4j.rl4j.observation.preprocessor.pooling
-
- cleanupPostTraining() - Method in class org.deeplearning4j.rl4j.learning.async.AsyncLearning
-
- clone() - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticCompGraph
-
- clone() - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticSeparate
-
- clone() - Method in interface org.deeplearning4j.rl4j.network.ac.IActorCritic
-
- clone() - Method in class org.deeplearning4j.rl4j.network.dqn.DQN
-
- clone() - Method in interface org.deeplearning4j.rl4j.network.dqn.IDQN
-
- clone() - Method in interface org.deeplearning4j.rl4j.network.NeuralNet
-
clone the Neural Net with the same paramaeters
- close() - Method in class org.deeplearning4j.rl4j.mdp.CartpoleNative
-
- close() - Method in class org.deeplearning4j.rl4j.mdp.toy.HardDeteministicToy
-
- close() - Method in class org.deeplearning4j.rl4j.mdp.toy.SimpleToy
-
- close() - Method in class org.deeplearning4j.rl4j.util.LegacyMDPWrapper
-
- close() - Method in class org.deeplearning4j.rl4j.util.VideoRecorder
-
Terminate the recording and close the video file
- codec(int) - Method in class org.deeplearning4j.rl4j.util.VideoRecorder.Builder
-
The codec to use for the video.
- computeGradient(INDArray, INDArray, IActivation, INDArray) - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticLoss
-
- computeGradientAndScore(INDArray, INDArray, IActivation, INDArray, boolean) - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticLoss
-
- computeScore(INDArray, INDArray, IActivation, INDArray, boolean) - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticLoss
-
- computeScoreArray(INDArray, INDArray, IActivation, INDArray) - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticLoss
-
- computeTarget(int, double, boolean) - Method in class org.deeplearning4j.rl4j.learning.sync.qlearning.discrete.TDTargetAlgorithm.BaseTDTargetAlgorithm
-
Compute the new estimated Q-Value for every transition in the batch
- computeTarget(int, double, boolean) - Method in class org.deeplearning4j.rl4j.learning.sync.qlearning.discrete.TDTargetAlgorithm.DoubleDQN
-
In litterature, this corresponds to:
Q(s_t, a_t) = R_{t+1} + \gamma * Q_{tar}(s_{t+1}, max_{a}Q(s_{t+1}, a))
- computeTarget(int, double, boolean) - Method in class org.deeplearning4j.rl4j.learning.sync.qlearning.discrete.TDTargetAlgorithm.StandardDQN
-
In litterature, this corresponds to:
Q(s_t, a_t) = R_{t+1} + \gamma * max_{a}Q_{tar}(s_{t+1}, a)
- computeTDTargets(List<Transition<Integer>>) - Method in class org.deeplearning4j.rl4j.learning.sync.qlearning.discrete.TDTargetAlgorithm.BaseTDTargetAlgorithm
-
- computeTDTargets(List<Transition<A>>) - Method in interface org.deeplearning4j.rl4j.learning.sync.qlearning.discrete.TDTargetAlgorithm.ITDTargetAlgorithm
-
Compute the updated estimated Q-Values for every transition
- concat(INDArray[]) - Static method in class org.deeplearning4j.rl4j.learning.sync.Transition
-
concat an array history into a single INDArry of as many channel
as element in the history array
- conf - Variable in class org.deeplearning4j.rl4j.learning.async.a3c.discrete.A3CThreadDiscrete
-
- conf - Variable in class org.deeplearning4j.rl4j.learning.async.nstep.discrete.AsyncNStepQLearningThreadDiscrete
-
- configuration - Variable in class org.deeplearning4j.rl4j.learning.async.a3c.discrete.A3CDiscrete
-
- configuration - Variable in class org.deeplearning4j.rl4j.learning.async.nstep.discrete.AsyncNStepQLearningDiscrete
-
- Configuration() - Constructor for class org.deeplearning4j.rl4j.learning.IHistoryProcessor.Configuration
-
- Configuration() - Constructor for class org.deeplearning4j.rl4j.network.ac.ActorCriticFactoryCompGraphStdConv.Configuration
-
- Configuration() - Constructor for class org.deeplearning4j.rl4j.network.ac.ActorCriticFactoryCompGraphStdDense.Configuration
-
- Configuration() - Constructor for class org.deeplearning4j.rl4j.network.ac.ActorCriticFactorySeparateStdDense.Configuration
-
- Configuration() - Constructor for class org.deeplearning4j.rl4j.network.dqn.DQNFactoryStdConv.Configuration
-
- Configuration() - Constructor for class org.deeplearning4j.rl4j.network.dqn.DQNFactoryStdDense.Configuration
-
- Constants - Class in org.deeplearning4j.rl4j.util
-
- Constants() - Constructor for class org.deeplearning4j.rl4j.util.Constants
-
- copy(NN) - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticCompGraph
-
- copy(NN) - Method in class org.deeplearning4j.rl4j.network.ac.ActorCriticSeparate
-
- copy(NN) - Method in interface org.deeplearning4j.rl4j.network.ac.IActorCritic
-
- copy(NN) - Method in class org.deeplearning4j.rl4j.network.dqn.DQN
-
- copy(NN) - Method in interface org.deeplearning4j.rl4j.network.dqn.IDQN
-
- copy(NN) - Method in interface org.deeplearning4j.rl4j.network.NeuralNet
-
copy the parameters from a neural net
- createFrame(byte[]) - Method in class org.deeplearning4j.rl4j.util.VideoRecorder
-
Create a VideoFrame from a byte array.
- createFrame(byte[], int, int) - Method in class org.deeplearning4j.rl4j.util.VideoRecorder
-
Create a VideoFrame from a byte array with different height and width than the video
the frame will need to be cropped or resized before being added to the video)
- createFrame(Pointer) - Method in class org.deeplearning4j.rl4j.util.VideoRecorder
-
Create a VideoFrame from a Pointer (to use for example with a INDarray).
- createFrame(Pointer, int, int) - Method in class org.deeplearning4j.rl4j.util.VideoRecorder
-
Create a VideoFrame from a Pointer with different height and width than the video
the frame will need to be cropped or resized before being added to the video)
- createSubdir() - Method in class org.deeplearning4j.rl4j.util.DataManager
-
- crop(int, int, int, int) - Method in class org.deeplearning4j.rl4j.util.VideoRecorder.VideoFrame
-
Crop the video to a specified size