Index

A B C D E F G H I L M N O P R S T U V W X Y Z 
All Classes and Interfaces|All Packages|Constant Field Values

T

tanh(Tensor) - Method in class io.github.kirstenali.deepj.tensor.cpu.CpuBackend
 
tanh(Tensor) - Method in class io.github.kirstenali.deepj.tensor.metal.MetalBackend
 
tanh(Tensor) - Method in interface io.github.kirstenali.deepj.tensor.TensorBackend
 
Tanh - Class in io.github.kirstenali.deepj.activations
 
Tanh() - Constructor for class io.github.kirstenali.deepj.activations.Tanh
 
tanhActivation() - Method in class io.github.kirstenali.deepj.tensor.Tensor
 
tanhInPlace() - Method in class io.github.kirstenali.deepj.tensor.Tensor
 
tanhInPlace(Tensor) - Method in class io.github.kirstenali.deepj.tensor.cpu.CpuBackend
 
tanhInPlace(Tensor) - Method in interface io.github.kirstenali.deepj.tensor.TensorBackend
 
Tensor - Class in io.github.kirstenali.deepj.tensor
 
Tensor(double[][]) - Constructor for class io.github.kirstenali.deepj.tensor.Tensor
 
Tensor(int, int) - Constructor for class io.github.kirstenali.deepj.tensor.Tensor
 
TensorAdapters - Class in io.github.kirstenali.deepj.tensor
Utilities for converting between Tensor (double[][]) and flat float[] arrays used by GPU runtimes.
TensorBackend - Interface in io.github.kirstenali.deepj.tensor
 
text() - Method in record class io.github.kirstenali.deepj.chatui.ChatMessage
Returns the value of the text record component.
TextDataset - Class in io.github.kirstenali.deepj.data
Simple in-memory dataset that samples random contiguous chunks from token ids.
TextDataset(int[], int, long) - Constructor for class io.github.kirstenali.deepj.data.TextDataset
 
TextGenerator - Class in io.github.kirstenali.deepj.models.gpt
Simple autoregressive text generation for GPTModel.
toIntTargets(Tensor) - Static method in class io.github.kirstenali.deepj.loss.CrossEntropyLoss
Converts a [n x 1] Tensor of class indices into an int[].
Tokenizer - Interface in io.github.kirstenali.deepj.tokenizers
 
tokenKeyToId() - Method in record class io.github.kirstenali.deepj.tokenizers.bpe.BPEModel
Returns the value of the tokenKeyToId record component.
TokenPair - Record Class in io.github.kirstenali.deepj.tokenizers.bpe
 
TokenPair(int, int) - Constructor for record class io.github.kirstenali.deepj.tokenizers.bpe.TokenPair
Creates an instance of a TokenPair record class.
toString() - Method in record class io.github.kirstenali.deepj.chatui.ChatMessage
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.data.Batch
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.models.gpt.GPTConfig
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.tokenizers.bpe.BPEModel
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.tokenizers.bpe.BPETokenizer
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.tokenizers.bpe.TokenPair
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.training.TrainingResult
Returns a string representation of this record class.
toString() - Method in record class io.github.kirstenali.deepj.transformer.TransformerStack
Returns a string representation of this record class.
train(int, int, int, double, Double) - Method in class io.github.kirstenali.deepj.training.Trainer
Train until maxSteps or until EMA loss goes below targetEmaLoss (if provided).
train(int, int, int, double, Double, int) - Method in class io.github.kirstenali.deepj.training.Trainer
 
train(int, int, int, double, Double, int, Trainer.StepHook) - Method in class io.github.kirstenali.deepj.training.Trainer
Train until maxSteps or until EMA loss goes below targetEmaLoss (if provided).
train(int, int, int, double, Double, Trainer.StepHook) - Method in class io.github.kirstenali.deepj.training.Trainer
Train until maxSteps or until EMA loss goes below targetEmaLoss (if provided).
train(String, int) - Method in class io.github.kirstenali.deepj.tokenizers.bpe.BPETrainer
 
Trainable - Interface in io.github.kirstenali.deepj.training
 
TrainClassicFNN - Class in io.github.kirstenali.deepj.examples
Example: classic MLP training using Linear + activation (via FNN) and the unified Trainer wrapper.
TrainClassicFNN() - Constructor for class io.github.kirstenali.deepj.examples.TrainClassicFNN
 
trainer(Layer, LossFunction, ParameterOptimizer, Tensor, Tensor, long) - Static method in class io.github.kirstenali.deepj.training.SupervisedTraining
 
trainer(GPTModel, TextDataset, double) - Static method in class io.github.kirstenali.deepj.training.CausalLMTraining
 
Trainer - Class in io.github.kirstenali.deepj.training
A small, reusable training loop wrapper.
Trainer(Trainer.StepFunction) - Constructor for class io.github.kirstenali.deepj.training.Trainer
 
Trainer.StepFunction - Interface in io.github.kirstenali.deepj.training
 
Trainer.StepHook - Interface in io.github.kirstenali.deepj.training
 
trainFromFile(Path, int) - Method in class io.github.kirstenali.deepj.tokenizers.bpe.BPETrainer
 
TrainingResult - Record Class in io.github.kirstenali.deepj.training
 
TrainingResult(int, double, double) - Constructor for record class io.github.kirstenali.deepj.training.TrainingResult
Creates an instance of a TrainingResult record class.
TrainSmallGPT - Class in io.github.kirstenali.deepj.examples
Example: tiny GPT training on a small text file using byte-level tokens.
TrainSmallGPT() - Constructor for class io.github.kirstenali.deepj.examples.TrainSmallGPT
 
trainStep(int) - Method in interface io.github.kirstenali.deepj.training.Trainer.StepFunction
Runs one optimization step and returns the average loss for that step.
trainStep(int) - Method in class io.github.kirstenali.deepj.training.Trainer
 
trainTokenizer(String, int) - Method in class io.github.kirstenali.deepj.tokenizers.bpe.BPETrainer
 
trainTokenizerFromFile(Path, int) - Method in class io.github.kirstenali.deepj.tokenizers.bpe.BPETrainer
 
TransformerBlock - Class in io.github.kirstenali.deepj.layers.transformer
Pre-LN Transformer block: x = x + Attn(LN(x)) x = x + MLP(LN(x))
TransformerBlock(int, int, int, Supplier<ActivationFunction>, Random) - Constructor for class io.github.kirstenali.deepj.layers.transformer.TransformerBlock
 
TransformerBlock(int, int, int, Random) - Constructor for class io.github.kirstenali.deepj.layers.transformer.TransformerBlock
 
TransformerBuilder - Class in io.github.kirstenali.deepj.transformer
Convenience builder for assembling transformer stacks.
TransformerBuilder() - Constructor for class io.github.kirstenali.deepj.transformer.TransformerBuilder
 
TransformerStack - Record Class in io.github.kirstenali.deepj.transformer
A simple sequential stack of TransformerBlocks.
TransformerStack(List<TransformerBlock>) - Constructor for record class io.github.kirstenali.deepj.transformer.TransformerStack
Creates an instance of a TransformerStack record class.
transpose() - Method in class io.github.kirstenali.deepj.tensor.Tensor
 
transpose(Tensor) - Method in class io.github.kirstenali.deepj.tensor.cpu.CpuBackend
 
transpose(Tensor) - Method in class io.github.kirstenali.deepj.tensor.metal.MetalBackend
 
transpose(Tensor) - Method in interface io.github.kirstenali.deepj.tensor.TensorBackend
 
A B C D E F G H I L M N O P R S T U V W X Y Z 
All Classes and Interfaces|All Packages|Constant Field Values