Uses of Class
io.github.kirstenali.deepj.tensor.Tensor

Packages that use Tensor

Package	Description
io.github.kirstenali.deepj.activations
io.github.kirstenali.deepj.autograd
io.github.kirstenali.deepj.inference
io.github.kirstenali.deepj.layers
io.github.kirstenali.deepj.layers.transformer
io.github.kirstenali.deepj.layers.transformer.attention
io.github.kirstenali.deepj.layers.transformer.norm
io.github.kirstenali.deepj.loss
io.github.kirstenali.deepj.nn	Neural networks composed from DeepJ tensors, layers, autograd and backend operations.
io.github.kirstenali.deepj.nn.transformer
io.github.kirstenali.deepj.nn.transformer.origin
io.github.kirstenali.deepj.optimisers
io.github.kirstenali.deepj.tensor
io.github.kirstenali.deepj.tensor.cpu
io.github.kirstenali.deepj.training
io.github.kirstenali.deepj.transformer
io.github.kirstenali.deepj.transformer.embeddings

Uses of Tensor in io.github.kirstenali.deepj.activations

Methods in io.github.kirstenali.deepj.activations that return Tensor

Modifier and Type	Method	Description
Tensor	ActivationFunction.forward(Tensor input)
Tensor	GELU.forward(Tensor input)
Tensor	ReLU.forward(Tensor input)
Tensor	Sigmoid.forward(Tensor input)
Tensor	SiLU.forward(Tensor input)
Tensor	Softmax.forward(Tensor input)
Tensor	Tanh.forward(Tensor input)
Tensor	Softmax.forwardCausal(Tensor logits, int sequenceLength, float scale)

Methods in io.github.kirstenali.deepj.activations with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	ActivationFunction.forward(Tensor input)
Tensor	GELU.forward(Tensor input)
Tensor	ReLU.forward(Tensor input)
Tensor	Sigmoid.forward(Tensor input)
Tensor	SiLU.forward(Tensor input)
Tensor	Softmax.forward(Tensor input)
Tensor	Tanh.forward(Tensor input)
Tensor	Softmax.forwardCausal(Tensor logits, int sequenceLength, float scale)

Uses of Tensor in io.github.kirstenali.deepj.autograd

Methods in io.github.kirstenali.deepj.autograd that return Tensor

Modifier and Type	Method	Description
static Tensor	BinaryOps.add(Tensor a, Tensor b)
static Tensor	BroadcastOps.add(Tensor a, Tensor b, boolean columns)
Tensor[]	BackwardRule.apply(Tensor gradient)
static Tensor	LayoutOps.batchedMatmul(Tensor a, Tensor b, int batches, boolean ta, boolean tb)
static Tensor	ViewOps.cast(Tensor input, DType dtype)
static Tensor	AttentionOps.causal(Tensor query, Tensor key, Tensor value, int heads, float scale)
static Tensor	AttentionOps.causal(Tensor query, Tensor key, Tensor value, int queryHeads, int keyValueHeads, float scale)
static Tensor	ActivationOps.causalSoftmax(Tensor input, int length, float scale)
static Tensor	SelectionOps.clamp(Tensor input, float minimum, float maximum)	The derivative is zero at either clipping boundary.
static Tensor	ReductionOps.columns(Tensor input)
static Tensor	ViewOps.contiguous(Tensor input)
static Tensor	ViewOps.copy(Tensor input)
static Tensor	BinaryOps.divide(Tensor a, Tensor b)
static Tensor	BroadcastOps.divide(Tensor a, Tensor b)
static Tensor	UnaryOps.divide(Tensor input, float value)
static Tensor	UnaryOps.exp(Tensor input)
static Tensor	ViewOps.expand(Tensor input, int... dimensions)
static Tensor	LayoutOps.gather(Tensor input, int[] indices)
static Tensor	ActivationOps.gelu(Tensor input)
Tensor	GradientState.gradient()
static Tensor	LayoutOps.indexSelect(Tensor input, int axis, int[] indices)
static Tensor	NormalizationOps.layerNorm(Tensor input, Tensor gamma, Tensor beta, float epsilon)
static Tensor	UnaryOps.log(Tensor input)
static Tensor	BinaryOps.matmul(Tensor a, Tensor b)
static Tensor	AxisReductionOps.maximum(Tensor input, int[] axes, boolean keep)
static Tensor	SelectionOps.maximum(Tensor input)	Tied maxima share the row gradient equally.
static Tensor	AxisReductionOps.mean(Tensor input, int[] axes, boolean keep)
static Tensor	ReductionOps.mean(Tensor input)
static Tensor	LayoutOps.mergeHeads(Tensor input, int heads)
static Tensor	BinaryOps.multiply(Tensor a, Tensor b)
static Tensor	BroadcastOps.multiply(Tensor a, Tensor b, boolean columns)
static Tensor	UnaryOps.neg(Tensor input)
static Tensor	ViewOps.permute(Tensor input, int... axes)
static Tensor	UnaryOps.pow(Tensor input, float exponent)
static Tensor	Autograd.record(Tensor output, BackwardRule rule, Tensor... inputs)
static Tensor	Autograd.recordSaved(Tensor output, BackwardRule rule, Tensor[] inputs, Tensor... saved)
static Tensor	ActivationOps.relu(Tensor input)
static Tensor	ViewOps.reshape(Tensor input, int... dimensions)
static Tensor	NormalizationOps.rmsNorm(Tensor input, Tensor gamma, float epsilon)
static Tensor	LayoutOps.rotary(Tensor input, Tensor cosine, Tensor sine, int length, boolean inverse)
static Tensor	ReductionOps.rows(Tensor input)
static Tensor	UnaryOps.scale(Tensor input, float value)
static Tensor	ViewOps.select(Tensor input, int dimension, int index)
static Tensor	UnaryOps.shift(Tensor input, float value)
static Tensor	ActivationOps.sigmoid(Tensor input)
static Tensor	ViewOps.slice(Tensor input, int dimension, int start, int stop, int step)
static Tensor	ActivationOps.softmax(Tensor input)
static Tensor	AxisReductionOps.softmax(Tensor input, int axis)
static Tensor	LayoutOps.splitHeads(Tensor input, int heads)
static Tensor	UnaryOps.sqrt(Tensor input)
static Tensor	BinaryOps.subtract(Tensor a, Tensor b)
static Tensor	BroadcastOps.subtract(Tensor a, Tensor b)
static Tensor	AxisReductionOps.sum(Tensor input, int[] axes, boolean keep)
static Tensor	ReductionOps.sum(Tensor input)
static Tensor	ActivationOps.swiGlu(Tensor gate, Tensor up)
static Tensor	ActivationOps.tanh(Tensor input)
static Tensor	UnaryOps.transpose(Tensor input)
static Tensor	ViewOps.transpose(Tensor input, int first, int second)
static Tensor	ViewOps.unsqueeze(Tensor input, int axis)
static Tensor	AxisReductionOps.variance(Tensor input, int[] axes, boolean keep, int correction)
static Tensor	ReductionOps.variance(Tensor input)
static Tensor	ViewOps.view(Tensor input, int... dimensions)

Methods in io.github.kirstenali.deepj.autograd with parameters of type Tensor

Modifier and Type	Method	Description
static Tensor	BinaryOps.add(Tensor a, Tensor b)
static Tensor	BroadcastOps.add(Tensor a, Tensor b, boolean columns)
Tensor[]	BackwardRule.apply(Tensor gradient)
static void	Autograd.backward(Tensor output)
static void	Autograd.backward(Tensor output, Tensor gradient)
static Tensor	LayoutOps.batchedMatmul(Tensor a, Tensor b, int batches, boolean ta, boolean tb)
void	GradientState.bindGradient(Tensor value)
static Tensor	ViewOps.cast(Tensor input, DType dtype)
static Tensor	AttentionOps.causal(Tensor query, Tensor key, Tensor value, int heads, float scale)
static Tensor	AttentionOps.causal(Tensor query, Tensor key, Tensor value, int queryHeads, int keyValueHeads, float scale)
static Tensor	ActivationOps.causalSoftmax(Tensor input, int length, float scale)
static Tensor	SelectionOps.clamp(Tensor input, float minimum, float maximum)	The derivative is zero at either clipping boundary.
static Tensor	ReductionOps.columns(Tensor input)
static Tensor	ViewOps.contiguous(Tensor input)
static Tensor	ViewOps.copy(Tensor input)
static Tensor	BinaryOps.divide(Tensor a, Tensor b)
static Tensor	BroadcastOps.divide(Tensor a, Tensor b)
static Tensor	UnaryOps.divide(Tensor input, float value)
static Tensor	UnaryOps.exp(Tensor input)
static Tensor	ViewOps.expand(Tensor input, int... dimensions)
static Tensor	LayoutOps.gather(Tensor input, int[] indices)
static Tensor	ActivationOps.gelu(Tensor input)
static Tensor	LayoutOps.indexSelect(Tensor input, int axis, int[] indices)
static Tensor	NormalizationOps.layerNorm(Tensor input, Tensor gamma, Tensor beta, float epsilon)
static Tensor	UnaryOps.log(Tensor input)
static Tensor	BinaryOps.matmul(Tensor a, Tensor b)
static Tensor	AxisReductionOps.maximum(Tensor input, int[] axes, boolean keep)
static Tensor	SelectionOps.maximum(Tensor input)	Tied maxima share the row gradient equally.
static Tensor	AxisReductionOps.mean(Tensor input, int[] axes, boolean keep)
static Tensor	ReductionOps.mean(Tensor input)
static Tensor	LayoutOps.mergeHeads(Tensor input, int heads)
static Tensor	BinaryOps.multiply(Tensor a, Tensor b)
static Tensor	BroadcastOps.multiply(Tensor a, Tensor b, boolean columns)
static Tensor	UnaryOps.neg(Tensor input)
static Tensor	ViewOps.permute(Tensor input, int... axes)
static Tensor	UnaryOps.pow(Tensor input, float exponent)
static Tensor	Autograd.record(Tensor output, BackwardRule rule, Tensor... inputs)
static Tensor	Autograd.recordSaved(Tensor output, BackwardRule rule, Tensor[] inputs, Tensor... saved)
static Tensor	ActivationOps.relu(Tensor input)
static Tensor	ViewOps.reshape(Tensor input, int... dimensions)
static Tensor	NormalizationOps.rmsNorm(Tensor input, Tensor gamma, float epsilon)
static Tensor	LayoutOps.rotary(Tensor input, Tensor cosine, Tensor sine, int length, boolean inverse)
static Tensor	ReductionOps.rows(Tensor input)
static Tensor	UnaryOps.scale(Tensor input, float value)
static Tensor	ViewOps.select(Tensor input, int dimension, int index)
static Tensor	UnaryOps.shift(Tensor input, float value)
static Tensor	ActivationOps.sigmoid(Tensor input)
static Tensor	ViewOps.slice(Tensor input, int dimension, int start, int stop, int step)
static Tensor	ActivationOps.softmax(Tensor input)
static Tensor	AxisReductionOps.softmax(Tensor input, int axis)
static Tensor	LayoutOps.splitHeads(Tensor input, int heads)
static Tensor	UnaryOps.sqrt(Tensor input)
static Tensor	BinaryOps.subtract(Tensor a, Tensor b)
static Tensor	BroadcastOps.subtract(Tensor a, Tensor b)
static Tensor	AxisReductionOps.sum(Tensor input, int[] axes, boolean keep)
static Tensor	ReductionOps.sum(Tensor input)
static Tensor	ActivationOps.swiGlu(Tensor gate, Tensor up)
static Tensor	ActivationOps.tanh(Tensor input)
static boolean	Autograd.tracks(Tensor... inputs)
static Tensor	UnaryOps.transpose(Tensor input)
static Tensor	ViewOps.transpose(Tensor input, int first, int second)
static Tensor	ViewOps.unsqueeze(Tensor input, int axis)
static Tensor	AxisReductionOps.variance(Tensor input, int[] axes, boolean keep, int correction)
static Tensor	ReductionOps.variance(Tensor input)
static Tensor	ViewOps.view(Tensor input, int... dimensions)

Uses of Tensor in io.github.kirstenali.deepj.inference

Methods in io.github.kirstenali.deepj.inference that return Tensor

Modifier and Type	Method	Description
Tensor	InferenceContext.attend(Object owner, Tensor query, Tensor key, Tensor value, int queryHeads, int keyValueHeads, float scale)
Tensor	InferenceContext.rotate(RotaryEmbedding rotary, Tensor input, int length, int heads)

Methods in io.github.kirstenali.deepj.inference with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	InferenceContext.attend(Object owner, Tensor query, Tensor key, Tensor value, int queryHeads, int keyValueHeads, float scale)
Tensor	InferenceContext.rotate(RotaryEmbedding rotary, Tensor input, int length, int heads)

Uses of Tensor in io.github.kirstenali.deepj.layers

Methods in io.github.kirstenali.deepj.layers that return Tensor

Modifier and Type	Method	Description
static Tensor	ProjectionOps.affine(Tensor input, Tensor weight, Tensor bias)
Tensor	BiasFreeLinear.forward(Tensor value)
Tensor	Layer.forward(Tensor input)
default Tensor	Layer.forward(Tensor input, InferenceContext context)
Tensor	Linear.forward(Tensor x)
Tensor	TiedEmbeddingProjection.forward(Tensor value)
static Tensor	ProjectionOps.linear(Tensor input, Tensor weight)
static Tensor	ProjectionOps.tied(Tensor input, Tensor embedding)

Methods in io.github.kirstenali.deepj.layers with parameters of type Tensor

Modifier and Type	Method	Description
static Tensor	ProjectionOps.affine(Tensor input, Tensor weight, Tensor bias)
Tensor	BiasFreeLinear.forward(Tensor value)
Tensor	Layer.forward(Tensor input)
default Tensor	Layer.forward(Tensor input, InferenceContext context)
Tensor	Linear.forward(Tensor x)
Tensor	TiedEmbeddingProjection.forward(Tensor value)
static Tensor	ProjectionOps.linear(Tensor input, Tensor weight)
static Tensor	ProjectionOps.tied(Tensor input, Tensor embedding)

Uses of Tensor in io.github.kirstenali.deepj.layers.transformer

Methods in io.github.kirstenali.deepj.layers.transformer that return Tensor

Modifier and Type	Method	Description
Tensor	SwiGLULayer.forward(Tensor x)

Methods in io.github.kirstenali.deepj.layers.transformer with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	SwiGLULayer.forward(Tensor x)

Uses of Tensor in io.github.kirstenali.deepj.layers.transformer.attention

Methods in io.github.kirstenali.deepj.layers.transformer.attention that return Tensor

Modifier and Type	Method	Description
Tensor	GroupedQueryAttention.forward(Tensor input)
Tensor	GroupedQueryAttention.forward(Tensor input, InferenceContext context)
Tensor	MultiHeadLatentAttention.forward(Tensor input)
Tensor	MultiHeadLatentAttention.forward(Tensor input, InferenceContext context)
Tensor	MultiHeadSelfAttention.forward(Tensor x)
Tensor	MultiHeadSelfAttention.forward(Tensor input, InferenceContext context)
protected Tensor	MultiHeadSelfAttention.transformQueryKey(Tensor heads, int seqLen)
protected Tensor	MultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, int offset)
protected Tensor	MultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, InferenceContext context)
protected Tensor	RoPEMultiHeadSelfAttention.transformQueryKey(Tensor heads, int seqLen)
protected Tensor	RoPEMultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, int offset)
protected Tensor	RoPEMultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, InferenceContext context)

Methods in io.github.kirstenali.deepj.layers.transformer.attention with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	GroupedQueryAttention.forward(Tensor input)
Tensor	GroupedQueryAttention.forward(Tensor input, InferenceContext context)
Tensor	MultiHeadLatentAttention.forward(Tensor input)
Tensor	MultiHeadLatentAttention.forward(Tensor input, InferenceContext context)
Tensor	MultiHeadSelfAttention.forward(Tensor x)
Tensor	MultiHeadSelfAttention.forward(Tensor input, InferenceContext context)
protected Tensor	MultiHeadSelfAttention.transformQueryKey(Tensor heads, int seqLen)
protected Tensor	MultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, int offset)
protected Tensor	MultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, InferenceContext context)
protected Tensor	RoPEMultiHeadSelfAttention.transformQueryKey(Tensor heads, int seqLen)
protected Tensor	RoPEMultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, int offset)
protected Tensor	RoPEMultiHeadSelfAttention.transformQueryKey(Tensor heads, int length, InferenceContext context)

Uses of Tensor in io.github.kirstenali.deepj.layers.transformer.norm

Methods in io.github.kirstenali.deepj.layers.transformer.norm that return Tensor

Modifier and Type	Method	Description
Tensor	LayerNorm1D.forward(Tensor input)
Tensor	RMSNorm1D.forward(Tensor input)

Methods in io.github.kirstenali.deepj.layers.transformer.norm with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	LayerNorm1D.forward(Tensor input)
Tensor	RMSNorm1D.forward(Tensor input)

Uses of Tensor in io.github.kirstenali.deepj.loss

Methods in io.github.kirstenali.deepj.loss that return Tensor

Modifier and Type	Method	Description
static Tensor	CrossEntropyLoss.forward(Tensor logits, int[] targets, boolean[] mask)
Tensor	CrossEntropyLoss.forward(Tensor predicted, Tensor actual)
Tensor	LossFunction.forward(Tensor predicted, Tensor actual)
Tensor	MSELoss.forward(Tensor predicted, Tensor actual)
static Tensor	CrossEntropyLoss.fromIntTargets(int[] targets)
static Tensor	CrossEntropyLoss.gradient(Tensor logits, int[] targets)
static Tensor	CrossEntropyLoss.gradient(Tensor logits, int[] targets, boolean[] mask)
Tensor	CrossEntropyLoss.gradient(Tensor predicted, Tensor actual)
Tensor	LossFunction.gradient(Tensor predicted, Tensor actual)
Tensor	MSELoss.gradient(Tensor predicted, Tensor actual)

Methods in io.github.kirstenali.deepj.loss with parameters of type Tensor

Modifier and Type	Method	Description
static Tensor	CrossEntropyLoss.forward(Tensor logits, int[] targets, boolean[] mask)
Tensor	CrossEntropyLoss.forward(Tensor predicted, Tensor actual)
Tensor	LossFunction.forward(Tensor predicted, Tensor actual)
Tensor	MSELoss.forward(Tensor predicted, Tensor actual)
static Tensor	CrossEntropyLoss.gradient(Tensor logits, int[] targets)
static Tensor	CrossEntropyLoss.gradient(Tensor logits, int[] targets, boolean[] mask)
Tensor	CrossEntropyLoss.gradient(Tensor predicted, Tensor actual)
Tensor	LossFunction.gradient(Tensor predicted, Tensor actual)
Tensor	MSELoss.gradient(Tensor predicted, Tensor actual)
static float	CrossEntropyLoss.loss(Tensor logits, int[] targets)
static float	CrossEntropyLoss.loss(Tensor logits, int[] targets, boolean[] mask)
float	CrossEntropyLoss.loss(Tensor predicted, Tensor actual)
float	LossFunction.loss(Tensor predicted, Tensor actual)
float	MSELoss.loss(Tensor predicted, Tensor actual)
static CrossEntropyResult	CrossEntropyLoss.result(Tensor logits, int[] targets)
static int[]	CrossEntropyLoss.toIntTargets(Tensor actual)

Uses of Tensor in io.github.kirstenali.deepj.nn

Methods in io.github.kirstenali.deepj.nn that return Tensor

Modifier and Type	Method	Description
Tensor	RecurrentState.cell()	Returns the value of the cell record component.
static Tensor	TensorJoin.concatenate(int axis, Tensor... inputs)
Tensor	Autoencoder.decode(Tensor encoded)
Tensor	Autoencoder.encode(Tensor input)
Tensor	Activation.forward(Tensor input)
Tensor	Autoencoder.forward(Tensor input)
Tensor	BatchNorm.forward(Tensor input)
Tensor	Conv1D.forward(Tensor input)
Tensor	Conv2D.forward(Tensor input)
Tensor	Dropout.forward(Tensor input)
Tensor	EncoderAttention.forward(Tensor input)
Tensor	EncoderAttention.forward(Tensor input, boolean[][] validTokens)
Tensor	Flatten.forward(Tensor input)
Tensor	FNN.forward(Tensor x)
Tensor	GlobalAveragePool.forward(Tensor input)
Tensor	Pool1D.forward(Tensor input)
Tensor	Pool2D.forward(Tensor input)
final Tensor	Recurrent.forward(Tensor input)
final Tensor	RecurrentCell.forward(Tensor input)
Tensor	Reshape.forward(Tensor input)
Tensor	Sequential.forward(Tensor input)
Tensor	TransformerEncoder.forward(Tensor input)
Tensor	TransformerEncoder.forward(Tensor input, boolean[][] validTokens)
Tensor	TransformerEncoderLayer.forward(Tensor input)
Tensor	TransformerEncoderLayer.forward(Tensor input, boolean[][] validTokens)
protected final Tensor	RecurrentCell.gate(Tensor projection, int index)
Tensor	RecurrentState.hidden()	Returns the value of the hidden record component.
Tensor	TransformerEncoder.meanPool(Tensor input, boolean[][] validTokens)
Tensor	BatchNorm.runningMean()
Tensor	BatchNorm.runningVariance()
Tensor	RecurrentResult.sequence()	Returns the value of the sequence record component.
static Tensor	TensorJoin.stack(int axis, Tensor... inputs)

Methods in io.github.kirstenali.deepj.nn that return types with arguments of type Tensor

Modifier and Type	Method	Description
List<Tensor>	BatchNorm.buffers()
List<Tensor>	Module.buffers()	Persistent non-parameter tensors, in deterministic traversal order.

Methods in io.github.kirstenali.deepj.nn with parameters of type Tensor

Modifier and Type	Method	Description
static Tensor	TensorJoin.concatenate(int axis, Tensor... inputs)
Tensor	Autoencoder.decode(Tensor encoded)
Tensor	Autoencoder.encode(Tensor input)
Tensor	Activation.forward(Tensor input)
Tensor	Autoencoder.forward(Tensor input)
Tensor	BatchNorm.forward(Tensor input)
Tensor	Conv1D.forward(Tensor input)
Tensor	Conv2D.forward(Tensor input)
Tensor	Dropout.forward(Tensor input)
Tensor	EncoderAttention.forward(Tensor input)
Tensor	EncoderAttention.forward(Tensor input, boolean[][] validTokens)
Tensor	Flatten.forward(Tensor input)
Tensor	FNN.forward(Tensor x)
Tensor	GlobalAveragePool.forward(Tensor input)
Tensor	Pool1D.forward(Tensor input)
Tensor	Pool2D.forward(Tensor input)
final Tensor	Recurrent.forward(Tensor input)
final Tensor	RecurrentCell.forward(Tensor input)
Tensor	Reshape.forward(Tensor input)
Tensor	Sequential.forward(Tensor input)
Tensor	TransformerEncoder.forward(Tensor input)
Tensor	TransformerEncoder.forward(Tensor input, boolean[][] validTokens)
Tensor	TransformerEncoderLayer.forward(Tensor input)
Tensor	TransformerEncoderLayer.forward(Tensor input, boolean[][] validTokens)
protected final Tensor	RecurrentCell.gate(Tensor projection, int index)
Tensor	TransformerEncoder.meanPool(Tensor input, boolean[][] validTokens)
protected RecurrentState	GRUCell.next(Tensor input, RecurrentState state)
protected RecurrentState	LSTMCell.next(Tensor input, RecurrentState state)
protected abstract RecurrentState	RecurrentCell.next(Tensor input, RecurrentState state)
protected RecurrentState	RNNCell.next(Tensor input, RecurrentState state)
final RecurrentResult	Recurrent.run(Tensor input, RecurrentState initial, int[] lengths)
static Tensor	TensorJoin.stack(int axis, Tensor... inputs)
final RecurrentState	RecurrentCell.step(Tensor input, RecurrentState state)
static Trainer	NdSupervisedTraining.trainer(Layer model, LossFunction loss, ParameterOptimizer optimizer, Tensor inputs, Tensor targets, long seed)

Constructors in io.github.kirstenali.deepj.nn with parameters of type Tensor

Modifier	Constructor	Description
	RecurrentResult(Tensor sequence, RecurrentState state)	Creates an instance of a RecurrentResult record class.
	RecurrentState(Tensor hidden)
	RecurrentState(Tensor hidden, Tensor cell)	Creates an instance of a RecurrentState record class.

Uses of Tensor in io.github.kirstenali.deepj.nn.transformer

Methods in io.github.kirstenali.deepj.nn.transformer that return Tensor

Modifier and Type	Method	Description
protected Tensor	DecoderOnlyModel.embed(int[] inputIds)
protected Tensor	DecoderOnlyModel.embed(int[] inputIds, int offset)
protected Tensor	DecoderOnlyModel.embedBatch(int[][] inputIds)
Tensor	CausalLM.forward(int[] inputIds)
Tensor	DecoderOnlyModel.forward(int[] inputIds)
default Tensor	CausalLM.forwardBatch(int[][] inputIds)
Tensor	DecoderOnlyModel.forwardBatch(int[][] inputIds)

Method parameters in io.github.kirstenali.deepj.nn.transformer with type arguments of type Tensor

Modifier and Type	Method	Description
static String	TextGenerator.generate(Function<int[],Tensor> forwarder, int maxSeqLen, Tokenizer tok, String prompt, int maxNewTokens, float temperature, int topK, long seed)
static String	TextGenerator.generate(Function<int[],Tensor> forwarder, int maxSeqLen, Tokenizer tok, String prompt, int maxNewTokens, float temperature, int topK, long seed, float repetitionPenalty)
static String	TextGenerator.generate(Function<int[],Tensor> forwarder, TransformerConfig cfg, Tokenizer tok, String prompt, int maxNewTokens, float temperature, int topK, long seed)
static String	TextGenerator.stream(Function<int[],Tensor> forwarder, int maxSeqLen, Tokenizer tok, String prompt, int maxNewTokens, float temperature, int topK, long seed, float repetitionPenalty, Consumer<String> listener)
static String	TextGenerator.stream(Function<int[],Tensor> forwarder, int maxSeqLen, Tokenizer tok, String prompt, int maxNewTokens, float temperature, int topK, long seed, Consumer<String> listener)

Uses of Tensor in io.github.kirstenali.deepj.nn.transformer.origin

Methods in io.github.kirstenali.deepj.nn.transformer.origin that return Tensor

Modifier and Type	Method	Description
protected Tensor	DeepJOrigin.embed(int[] inputIds)
protected Tensor	DeepJOrigin.embed(int[] inputIds, int offset)
protected Tensor	DeepJOrigin.embedBatch(int[][] inputIds)

Uses of Tensor in io.github.kirstenali.deepj.optimisers

Fields in io.github.kirstenali.deepj.optimisers declared as Tensor

Modifier and Type	Field	Description
final Tensor	Parameter.value

Methods in io.github.kirstenali.deepj.optimisers that return Tensor

Modifier and Type	Method	Description
Tensor	Parameter.grad()

Methods in io.github.kirstenali.deepj.optimisers that return types with arguments of type Tensor

Modifier and Type	Method	Description
List<Tensor>	AdamW.State.firstMoments()	Returns the value of the firstMoments record component.
Map<Integer,Tensor>	AdamW.State.masterWeights()	Returns the value of the masterWeights record component.
List<Tensor>	AdamW.State.secondMoments()	Returns the value of the secondMoments record component.

Methods in io.github.kirstenali.deepj.optimisers with parameters of type Tensor

Modifier and Type	Method	Description
void	Parameter.setGradient(Tensor gradient)

Constructors in io.github.kirstenali.deepj.optimisers with parameters of type Tensor

Modifier	Constructor	Description
	Parameter(Tensor value)
	Parameter(String name, Tensor value)

Constructor parameters in io.github.kirstenali.deepj.optimisers with type arguments of type Tensor

Modifier	Constructor	Description
	State(long step, float learningRate, float beta1, float beta2, float epsilon, float weightDecay, List<Tensor> firstMoments, List<Tensor> secondMoments)
	State(long step, float learningRate, float beta1, float beta2, float epsilon, float weightDecay, List<Tensor> firstMoments, List<Tensor> secondMoments, Map<Integer,Tensor> masterWeights)	Creates an instance of a State record class.
	State(long step, float learningRate, float beta1, float beta2, float epsilon, float weightDecay, List<Tensor> firstMoments, List<Tensor> secondMoments, Map<Integer,Tensor> masterWeights)	Creates an instance of a State record class.

Uses of Tensor in io.github.kirstenali.deepj.tensor

Methods in io.github.kirstenali.deepj.tensor that return Tensor

Modifier and Type	Method	Description
Tensor	CausalAttentionResult.accumulatedOutput()	Returns the value of the accumulatedOutput record component.
Tensor	Tensor.add(Tensor other)
Tensor	TensorBackend.add(Tensor a, Tensor b)
Tensor	Tensor.addBroadcastCols(Tensor colVector)
Tensor	TensorBackend.addBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.addBroadcastRows(Tensor rowVector)
Tensor	TensorBackend.addBroadcastRows(Tensor a, Tensor rowVector)
Tensor	Tensor.addInPlace(Tensor other)
Tensor	Tensor.addRowVector(Tensor rowVector)
Tensor	TensorBackend.addRowVector(Tensor a, Tensor rowVector)
Tensor	Tensor.addScalar(float s)
Tensor	TensorBackend.addScalar(Tensor a, float scalar)
Tensor	Tensor.addScalarInPlace(float scalar)
static Tensor	CpuNdMatmul.apply(Tensor left, Tensor right)
static Tensor	CpuNdSoftmax.apply(Tensor input, int axis)
static Tensor	CpuNdSwiGlu.apply(Tensor gate, Tensor up)
static Tensor	CpuNdNormalization.backward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma, boolean centered)
default Tensor	TensorBackend.batchedMatmul(Tensor left, Tensor right, int batches, boolean transposeLeft, boolean transposeRight)
static Tensor	CpuNdElementwise.binary(Tensor left, Tensor right, CpuNdElementwise.Binary operation)
static Tensor	CpuNdElementwise.cast(Tensor input, DType dtype)
default Tensor	TensorBackend.cast(Tensor input, DType dtype)
static Tensor	Tensor.causalMask(int size)
default Tensor	TensorBackend.causalMask(Tensor input, int sequenceLength)
default Tensor	TensorBackend.causalSoftmax(Tensor input, int sequenceLength, float scale)
Tensor	Tensor.clamp(float min, float max)
Tensor	TensorBackend.clamp(Tensor a, float min, float max)
final Tensor	TensorManipulations.contiguous()
Tensor	Tensor.copy()	Bit-preserving independent storage with an identity autograd edge.
default Tensor	TensorBackend.copy(Tensor source)	Independent storage with the same float bits, without recording an autograd operation.
final Tensor	TensorManipulations.copyFrom(Tensor source)	Independent value assignment with an explicit cast to this tensor's dtype; shapes must match.
Tensor	TensorBackend.crossEntropyGradient(Tensor logits, int[] targets)
Tensor	TensorBackend.crossEntropyGradient(Tensor logits, int[] targets, boolean[] mask)
Tensor	NormalizationResult.denominator()	Returns the value of the denominator record component.
Tensor	Tensor.detach()	Independent detached storage, preserving DeepJ's existing copy-based detach behavior.
Tensor	Tensor.divide(Tensor other)
Tensor	TensorBackend.divide(Tensor a, Tensor b)
Tensor	Tensor.divideBroadcastCols(Tensor colVector)
Tensor	TensorBackend.divideBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.divideInPlace(Tensor other)
Tensor	Tensor.divideScalar(float s)
Tensor	TensorBackend.divideScalar(Tensor a, float scalar)
Tensor	Tensor.divideScalarInPlace(float scalar)
Tensor	TensorScope.escape(Tensor tensor)
Tensor	Tensor.exp()
Tensor	TensorBackend.exp(Tensor a)
default Tensor	TensorBackend.expand(Tensor input, int... shape)
final Tensor	TensorManipulations.expand(int... shape)
Tensor	Tensor.expInPlace()
static Tensor	Tensor.from(float[] values)
static Tensor	Tensor.from(float[] values, int... shape)
static Tensor	Tensor.from(float[] values, DType dtype)
static Tensor	Tensor.from(float[] values, DType dtype, int... shape)
static Tensor	Tensor.from2D(float[][] data)
static Tensor	TensorAdapters.fromIntColumn(int[] values)
Tensor	SwiGluBackwardResult.gateGradient()	Returns the value of the gateGradient record component.
Tensor	TensorBackend.gelu(Tensor a)
Tensor	Tensor.geluActivation()
Tensor	TensorBackend.geluBackward(Tensor input, Tensor gradOutput)
Tensor	Tensor.geluInPlace()
Tensor	Tensor.getRow(int row)
Tensor	Tensor.grad()
static Tensor	CpuNdIndexing.gradient(Tensor gradient, int[] shape, int axis, int[] indices)
Tensor	CrossEntropyResult.gradient()	Returns the value of the gradient record component.
default Tensor	TensorBackend.headPrefixMatmul(Tensor left, Tensor cache, int heads, int length, boolean transposeCache)
final Tensor	TensorManipulations.indexAddInPlace(int axis, int[] indices, Tensor updates)
default Tensor	TensorBackend.indexSelect(Tensor input, int axis, int[] indices)
final Tensor	TensorManipulations.indexSelect(int axis, int... indices)
default Tensor	TensorBackend.indexSelectGradient(Tensor gradient, int[] shape, int axis, int[] indices)
Tensor	AttentionGradients.key()	Returns the value of the key record component.
Tensor	TensorBackend.layerNormBackward(Tensor dXHat, Tensor xHat, Tensor std, int dim)
Tensor	Tensor.log()
Tensor	TensorBackend.log(Tensor a)
Tensor	Tensor.logInPlace()
Tensor	CrossEntropyResult.loss()	Returns the value of the loss record component.
static Tensor	CpuNdPosition.mask(Tensor input, int length)
static Tensor	PackedLayouts.matmul(TensorBackend backend, Tensor left, Tensor right, int batches, boolean transposeLeft, boolean transposeRight)
Tensor	Tensor.matmul(Tensor other)
Tensor	TensorBackend.matmul(Tensor a, Tensor b)
Tensor	Tensor.maxAlongRows()
Tensor	TensorBackend.maxAlongRows(Tensor a)
final Tensor	TensorManipulations.maxAxes(int... axes)
static Tensor	CpuNdStatistics.maximum(Tensor input, int[] axes, boolean keep)
final Tensor	TensorManipulations.maximum(int axis, boolean keep)	Tied maxima share their gradient equally.
default Tensor	TensorBackend.maximumAxes(Tensor input, int[] axes, boolean keepDimensions)
default Tensor	TensorBackend.maximumAxesBackward(Tensor input, Tensor maximum, Tensor gradient, int[] axes, boolean keep)
static Tensor	CpuNdStatistics.maximumGradient(Tensor input, Tensor maximum, Tensor gradient, int[] axes, boolean keep)
static Tensor	CpuNdReduction.mean(Tensor input, int[] axes, boolean keep)
final Tensor	TensorManipulations.mean(int axis, boolean keep)
Tensor	Tensor.meanAlongRows()
Tensor	TensorBackend.meanAlongRows(Tensor a)
default Tensor	TensorBackend.meanAxes(Tensor input, int[] axes, boolean keepDimensions)
final Tensor	TensorManipulations.meanAxes(int... axes)
static Tensor	HeadLayout.merge(Tensor input, int heads)
static Tensor	PackedLayouts.merge(TensorBackend backend, Tensor input, int heads)
default Tensor	TensorBackend.mergeHeads(Tensor input, int heads)
Tensor	Tensor.multiply(Tensor other)
Tensor	TensorBackend.multiply(Tensor a, Tensor b)
Tensor	Tensor.multiplyBroadcastCols(Tensor colVector)
Tensor	TensorBackend.multiplyBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.multiplyBroadcastRows(Tensor rowVector)
Tensor	TensorBackend.multiplyBroadcastRows(Tensor a, Tensor rowVector)
Tensor	Tensor.multiplyInPlace(Tensor other)
Tensor	Tensor.multiplyScalar(float s)
Tensor	TensorBackend.multiplyScalar(Tensor a, float scalar)
Tensor	Tensor.multiplyScalarInPlace(float scalar)
Tensor	Tensor.neg()
Tensor	TensorBackend.neg(Tensor a)
Tensor	Tensor.negInPlace()
Tensor	CausalAttentionResult.normalization()	Returns the value of the normalization record component.
default Tensor	TensorBackend.normalizationBackward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma, boolean centered)
Tensor	NormalizationResult.normalized()	Returns the value of the normalized record component.
Tensor	RmsNormResult.normalized()	Returns the value of the normalized record component.
static Tensor	Tensor.ones(int rows, int cols)
static Tensor	TensorManipulations.ones(DType dtype, int... shape)
Tensor	CausalAttentionResult.output()	Returns the value of the output record component.
Tensor	NormalizationResult.output()	Returns the value of the output record component.
Tensor	RmsNormResult.output()	Returns the value of the output record component.
static Tensor	HeadLayout.pack(Tensor input)
default Tensor	TensorBackend.permute(Tensor input, int... axes)
final Tensor	TensorManipulations.permute(int... axes)
Tensor	Tensor.pow(float exponent)
Tensor	TensorBackend.pow(Tensor a, float exponent)
static Tensor	HeadCacheOps.prefix(TensorBackend backend, Tensor cache, int heads, int length)
Tensor	AttentionGradients.query()	Returns the value of the query record component.
static Tensor	Tensor.random(int rows, int cols, Random rand)
static Tensor	TensorManipulations.random(DType dtype, Random random, int... shape)
static Tensor	TensorEncoding.readValuesToBackend(DataInput input, DType dtype, int... shape)	Explicit checkpoint transfer; private host staging is released after an independent backend copy.
Tensor	TensorBackend.relu(Tensor a)
Tensor	Tensor.reluActivation()
Tensor	TensorBackend.reluBackward(Tensor input, Tensor gradOutput)
Tensor	Tensor.reluInPlace()
Tensor	Tensor.requiresGrad(boolean required)
default Tensor	TensorBackend.reshape(Tensor input, int... shape)
final Tensor	TensorManipulations.reshape(int... shape)
Tensor	Tensor.retainDeviceBuffer()
Tensor	RmsNormResult.rms()	Returns the value of the rms record component.
default Tensor	TensorBackend.rmsNormBackward(Tensor gradient, Tensor normalized, Tensor rms, Tensor gamma)
static Tensor	CpuNdPosition.rotary(Tensor input, Tensor cosine, Tensor sine, int length, boolean inverse)
default Tensor	TensorBackend.rotary(Tensor input, Tensor cosine, Tensor sine, int sequenceLength, boolean inverse)
static Tensor	CpuRowAccess.sample(Tensor input, int count, Random random)
static Tensor	Tensor.sampleRows(Tensor t, int n, Random rnd)
static Tensor	CpuNdIndexing.select(Tensor input, int axis, int[] indices)
static Tensor	CpuRowAccess.select(Tensor input, int[] rows, int columns)
default Tensor	TensorBackend.select(Tensor input, int axis, int index)
final Tensor	TensorManipulations.select(int axis, int index)
Tensor	TensorBackend.sigmoid(Tensor a)
Tensor	Tensor.sigmoidActivation()
Tensor	Tensor.sigmoidInPlace()
default Tensor	TensorBackend.slice(Tensor input, int axis, int start, int stop, int step)
final Tensor	TensorManipulations.slice(int axis, int start, int stop)
final Tensor	TensorManipulations.slice(int axis, int start, int stop, int step)
static Tensor	CpuNdReduction.sliceGradient(Tensor gradient, int[] dimensions, int axis, int start, int step)
default Tensor	TensorBackend.sliceGradient(Tensor gradient, int[] shape, int axis, int start, int step)
static Tensor	Tensor.sliceRows(Tensor t, int[] rowIndices, int cols)
default Tensor	TensorBackend.sliceRows(Tensor input, int[] rows)
static Tensor	CpuNdPosition.softmax(Tensor input, int length, float scale)
final Tensor	TensorManipulations.softmax(int axis)
default Tensor	TensorBackend.softmaxAxis(Tensor input, int axis)
default Tensor	TensorBackend.softmaxAxisBackward(Tensor gradient, Tensor output, int axis)
Tensor	TensorBackend.softmaxBackward(Tensor gradOutput, Tensor softmaxOut)
default Tensor	TensorBackend.softmaxBackward(Tensor gradOutput, Tensor softmaxOut, float scale)
Tensor	Tensor.softmaxRows()
Tensor	TensorBackend.softmaxRows(Tensor logits)
static Tensor	HeadLayout.split(Tensor input, int heads)
static Tensor	PackedLayouts.split(TensorBackend backend, Tensor input, int heads)
default Tensor	TensorBackend.splitHeads(Tensor input, int heads)
Tensor	Tensor.sqrt()
Tensor	TensorBackend.sqrt(Tensor a)
Tensor	Tensor.sqrtInPlace()
Tensor	Tensor.subtract(Tensor other)
Tensor	TensorBackend.subtract(Tensor a, Tensor b)
Tensor	Tensor.subtractBroadcastCols(Tensor colVector)
Tensor	TensorBackend.subtractBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.subtractInPlace(Tensor other)
static Tensor	CpuNdReduction.sum(Tensor input, int[] axes, boolean keep)
final Tensor	TensorManipulations.sum(int axis, boolean keepDimensions)
Tensor	Tensor.sumAlongCols()
Tensor	TensorBackend.sumAlongCols(Tensor a)
Tensor	Tensor.sumAlongRows()
Tensor	TensorBackend.sumAlongRows(Tensor a)
default Tensor	TensorBackend.sumAxes(Tensor input, int[] axes, boolean keepDimensions)
final Tensor	TensorManipulations.sumAxes(int... axes)
Tensor	Tensor.sumRows()
Tensor	TensorBackend.sumRows(Tensor a)
Tensor	Tensor.sumTensor()
default Tensor	TensorBackend.sumToShape(Tensor input, int... shape)
default Tensor	TensorBackend.swiGlu(Tensor gate, Tensor up)
Tensor	TensorBackend.tanh(Tensor a)
Tensor	Tensor.tanhActivation()
Tensor	Tensor.tanhInPlace()
protected Tensor	Tensor.tensor()
protected abstract Tensor	TensorManipulations.tensor()
final Tensor	TensorManipulations.to(DType dtype)
Tensor	Tensor.transpose()
Tensor	TensorBackend.transpose(Tensor a)
final Tensor	TensorManipulations.transpose(int first, int second)
static Tensor	CpuNdElementwise.unary(Tensor input, CpuNdElementwise.Unary operation)
static Tensor	TensorAdapters.unpackF32(float[] flat, int rows, int cols)
default Tensor	TensorBackend.unsqueeze(Tensor input, int axis)
final Tensor	TensorManipulations.unsqueeze(int axis)
Tensor	SwiGluBackwardResult.upGradient()	Returns the value of the upGradient record component.
Tensor	AttentionGradients.value()	Returns the value of the value record component.
static Tensor	CpuNdStatistics.variance(Tensor input, int[] axes, boolean keep, int correction)
final Tensor	TensorManipulations.variance(int axis, boolean keep)	Population variance; use the correction overload for a different denominator.
final Tensor	TensorManipulations.variance(int axis, boolean keep, int correction)
Tensor	Tensor.varianceAlongRows()
Tensor	TensorBackend.varianceAlongRows(Tensor a)
default Tensor	TensorBackend.varianceAxes(Tensor input, int[] axes, boolean keepDimensions, int correction)
final Tensor	TensorManipulations.varianceAxes(int[] axes, boolean keep, int correction)
default Tensor	TensorBackend.view(Tensor input, int... shape)
final Tensor	TensorManipulations.view(int... shape)
static Tensor	Tensor.zeros(int rows, int cols)
default Tensor	TensorBackend.zeros(int rows, int columns)	Backend-owned positive-zero storage; ordinary Tensor constructors stay CPU-backed.
default Tensor	TensorBackend.zeros(DType dtype, int... shape)
static Tensor	TensorManipulations.zeros(DType dtype, int... shape)

Methods in io.github.kirstenali.deepj.tensor with parameters of type Tensor

Modifier and Type	Method	Description
static void	Tensor.adamWUpdate(Tensor w, Tensor g, Tensor mt, Tensor vt, float lr, float beta1, float beta2, float eps, float weightDecay, float bc1, float bc2)
void	TensorBackend.adamWUpdate(Tensor w, Tensor g, Tensor mt, Tensor vt, float lr, float beta1, float beta2, float eps, float weightDecay, float bc1, float bc2)
static void	CpuNdIndexing.add(Tensor target, int axis, int[] indices, Tensor updates)
Tensor	Tensor.add(Tensor other)
Tensor	TensorBackend.add(Tensor a, Tensor b)
Tensor	Tensor.addBroadcastCols(Tensor colVector)
Tensor	TensorBackend.addBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.addBroadcastRows(Tensor rowVector)
Tensor	TensorBackend.addBroadcastRows(Tensor a, Tensor rowVector)
Tensor	Tensor.addInPlace(Tensor other)
void	TensorBackend.addInPlace(Tensor a, Tensor b)
static void	CpuNdIndexing.addMatrixRows(Tensor target, int[] indices, Tensor updates)
Tensor	Tensor.addRowVector(Tensor rowVector)
Tensor	TensorBackend.addRowVector(Tensor a, Tensor rowVector)
Tensor	TensorBackend.addScalar(Tensor a, float scalar)
void	TensorBackend.addScalarInPlace(Tensor a, float s)
static void	CpuNdAssignment.apply(Tensor target, Tensor source)
static CrossEntropyResult	CpuNdCrossEntropy.apply(Tensor logits, int[] targets, boolean[] mask)
static Tensor	CpuNdMatmul.apply(Tensor left, Tensor right)
static NormalizationResult	CpuNdNormalization.apply(Tensor input, Tensor gamma, Tensor beta, float epsilon, boolean centered)
static Tensor	CpuNdSoftmax.apply(Tensor input, int axis)
static Tensor	CpuNdSwiGlu.apply(Tensor gate, Tensor up)
static CrossEntropyResult	CrossEntropyDefaults.apply(TensorBackend backend, Tensor logits, int[] targets, boolean[] mask)
static void	CpuRowAccess.assign(Tensor target, int row, Tensor source, int sourceRow)
static Tensor	CpuNdNormalization.backward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma, boolean centered)
static SwiGluBackwardResult	CpuNdSwiGlu.backward(Tensor gradient, Tensor gate, Tensor up)
static void	NormalizationChecks.backward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma)
void	Tensor.backward(Tensor gradient)
default Tensor	TensorBackend.batchedMatmul(Tensor left, Tensor right, int batches, boolean transposeLeft, boolean transposeRight)
static int	SequenceLayout.batches(Tensor input)
static Tensor	CpuNdElementwise.binary(Tensor left, Tensor right, CpuNdElementwise.Binary operation)
static void	CpuNdElementwise.binaryInto(Tensor left, Tensor right, Tensor output, CpuNdElementwise.Binary operation)
static Tensor	CpuNdElementwise.cast(Tensor input, DType dtype)
default Tensor	TensorBackend.cast(Tensor input, DType dtype)
CausalAttentionResult	CausalAttentionBackend.causalAttention(Tensor query, Tensor key, Tensor value, int heads, float scale)
default CausalAttentionResult	CausalAttentionBackend.causalAttention(Tensor query, Tensor key, Tensor value, int queryHeads, int keyValueHeads, float scale)
AttentionGradients	CausalAttentionBackend.causalAttentionBackward(Tensor query, Tensor key, Tensor value, CausalAttentionResult saved, Tensor gradient, int heads, float scale)
default AttentionGradients	CausalAttentionBackend.causalAttentionBackward(Tensor query, Tensor key, Tensor value, CausalAttentionResult saved, Tensor gradient, int queryHeads, int keyValueHeads, float scale)
default Tensor	TensorBackend.causalMask(Tensor input, int sequenceLength)
default Tensor	TensorBackend.causalSoftmax(Tensor input, int sequenceLength, float scale)
Tensor	TensorBackend.clamp(Tensor a, float min, float max)
static int	CrossEntropyChecks.classes(Tensor logits)
default Tensor	TensorBackend.copy(Tensor source)	Independent storage with the same float bits, without recording an autograd operation.
final Tensor	TensorManipulations.copyFrom(Tensor source)	Independent value assignment with an explicit cast to this tensor's dtype; shapes must match.
default void	TensorBackend.copyInto(Tensor target, Tensor source)	Explicitly assigns logical values, casting to the destination's storage dtype.
default CrossEntropyResult	TensorBackend.crossEntropy(Tensor logits, int[] targets)
default CrossEntropyResult	TensorBackend.crossEntropy(Tensor logits, int[] targets, boolean[] mask)
Tensor	TensorBackend.crossEntropyGradient(Tensor logits, int[] targets)
Tensor	TensorBackend.crossEntropyGradient(Tensor logits, int[] targets, boolean[] mask)
float	TensorBackend.crossEntropyLoss(Tensor logits, int[] targets)
float	TensorBackend.crossEntropyLoss(Tensor logits, int[] targets, boolean[] mask)
Tensor	Tensor.divide(Tensor other)
Tensor	TensorBackend.divide(Tensor a, Tensor b)
Tensor	Tensor.divideBroadcastCols(Tensor colVector)
Tensor	TensorBackend.divideBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.divideInPlace(Tensor other)
void	TensorBackend.divideInPlace(Tensor a, Tensor b)
Tensor	TensorBackend.divideScalar(Tensor a, float scalar)
void	TensorBackend.divideScalarInPlace(Tensor a, float s)
Tensor	TensorScope.escape(Tensor tensor)
Tensor	TensorBackend.exp(Tensor a)
default Tensor	TensorBackend.expand(Tensor input, int... shape)
void	TensorBackend.expInPlace(Tensor a)
void	BackwardBuffers.expose(Tensor tensor)
static void	NormalizationChecks.forward(Tensor input, Tensor gamma, Tensor beta, float epsilon, boolean centered)
Tensor	TensorBackend.gelu(Tensor a)
Tensor	TensorBackend.geluBackward(Tensor input, Tensor gradOutput)
void	TensorBackend.geluInPlace(Tensor a)
static Tensor	CpuNdIndexing.gradient(Tensor gradient, int[] shape, int axis, int[] indices)
default Tensor	TensorBackend.headPrefixMatmul(Tensor left, Tensor cache, int heads, int length, boolean transposeCache)
default void	TensorBackend.indexAdd(Tensor target, int axis, int[] indices, Tensor updates)
final Tensor	TensorManipulations.indexAddInPlace(int axis, int[] indices, Tensor updates)
default Tensor	TensorBackend.indexSelect(Tensor input, int axis, int[] indices)
default Tensor	TensorBackend.indexSelectGradient(Tensor gradient, int[] shape, int axis, int[] indices)
Tensor	TensorBackend.layerNormBackward(Tensor dXHat, Tensor xHat, Tensor std, int dim)
static void	NormalizationChecks.layerStatistics(Tensor input, Tensor std)
static int	SequenceLayout.length(Tensor input)
Tensor	TensorBackend.log(Tensor a)
void	TensorBackend.logInPlace(Tensor a)
static float	CpuNdCrossEntropy.loss(Tensor logits, int[] targets, boolean[] mask)
static Tensor	CpuNdPosition.mask(Tensor input, int length)
default void	TensorBackend.materializeTensor(Tensor t)
static Tensor	PackedLayouts.matmul(TensorBackend backend, Tensor left, Tensor right, int batches, boolean transposeLeft, boolean transposeRight)
Tensor	Tensor.matmul(Tensor other)
Tensor	TensorBackend.matmul(Tensor a, Tensor b)
static boolean	CpuNdElementwise.matrix(Tensor input)
Tensor	TensorBackend.maxAlongRows(Tensor a)
static Tensor	CpuNdStatistics.maximum(Tensor input, int[] axes, boolean keep)
default Tensor	TensorBackend.maximumAxes(Tensor input, int[] axes, boolean keepDimensions)
default Tensor	TensorBackend.maximumAxesBackward(Tensor input, Tensor maximum, Tensor gradient, int[] axes, boolean keep)
static Tensor	CpuNdStatistics.maximumGradient(Tensor input, Tensor maximum, Tensor gradient, int[] axes, boolean keep)
static Tensor	CpuNdReduction.mean(Tensor input, int[] axes, boolean keep)
Tensor	TensorBackend.meanAlongRows(Tensor a)
default Tensor	TensorBackend.meanAxes(Tensor input, int[] axes, boolean keepDimensions)
static Tensor	HeadLayout.merge(Tensor input, int heads)
static Tensor	PackedLayouts.merge(TensorBackend backend, Tensor input, int heads)
default Tensor	TensorBackend.mergeHeads(Tensor input, int heads)
Tensor	Tensor.multiply(Tensor other)
Tensor	TensorBackend.multiply(Tensor a, Tensor b)
Tensor	Tensor.multiplyBroadcastCols(Tensor colVector)
Tensor	TensorBackend.multiplyBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.multiplyBroadcastRows(Tensor rowVector)
Tensor	TensorBackend.multiplyBroadcastRows(Tensor a, Tensor rowVector)
Tensor	Tensor.multiplyInPlace(Tensor other)
void	TensorBackend.multiplyInPlace(Tensor a, Tensor b)
Tensor	TensorBackend.multiplyScalar(Tensor a, float scalar)
void	TensorBackend.multiplyScalarInPlace(Tensor a, float s)
Tensor	TensorBackend.neg(Tensor a)
void	TensorBackend.negInPlace(Tensor a)
default Tensor	TensorBackend.normalizationBackward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma, boolean centered)
default NormalizationResult	TensorBackend.normalize(Tensor input, Tensor gamma, Tensor beta, float epsilon, boolean centered)
static PackedMatmulShape	PackedMatmulShape.of(Tensor left, Tensor right, int batches, boolean ta, boolean tb)
static DType	NormalizationChecks.outputType(Tensor input, Tensor gamma, Tensor beta, boolean centered)
static Tensor	HeadLayout.pack(Tensor input)
static float[]	TensorAdapters.packF32(Tensor t)
static void	NormalizationChecks.parameter(Tensor parameter, int width)
default Tensor	TensorBackend.permute(Tensor input, int... axes)
static int	CrossEntropyChecks.positions(Tensor logits)
Tensor	TensorBackend.pow(Tensor a, float exponent)
static Tensor	HeadCacheOps.prefix(TensorBackend backend, Tensor cache, int heads, int length)
static void	TensorEncoding.readInto(DataInput input, Tensor tensor)
static void	TensorEncoding.readValues(DataInput input, Tensor tensor)
Tensor	TensorBackend.relu(Tensor a)
Tensor	TensorBackend.reluBackward(Tensor input, Tensor gradOutput)
void	TensorBackend.reluInPlace(Tensor a)
static void	HeadLayout.require(Tensor input, int heads)
static void	MatrixShapeChecks.requireCausal(Tensor input, int sequenceLength)
void	TensorDescriptor.requireMatches(Tensor tensor)
static void	MatrixShapeChecks.requireMatrix(Tensor input)
static void	HeadCacheOps.requirePrefix(Tensor cache, int heads, int length)
static void	MatrixShapeChecks.requireRotary(Tensor input, Tensor cosine, Tensor sine, int sequenceLength)
static void	MatrixShapeChecks.requireRowMaximum(Tensor input)
static void	Tensor.requireSameShape(Tensor a, Tensor b, String op)
static void	OptimizerChecks.requireState(Tensor weight, Tensor gradient, Tensor first, Tensor second)
static void	CrossEntropyChecks.requireTargets(Tensor logits, int[] targets)
static void	Tensor.requireTargetsMatchRows(Tensor logits, int[] targets)
static void	SequenceLayout.requireWidth(Tensor input, int width)
static void	HeadCacheOps.requireWrite(Tensor cache, Tensor update, int heads, int offset)
default Tensor	TensorBackend.reshape(Tensor input, int... shape)
static void	TensorEncoding.restore(Tensor target, Tensor source)	Explicit host-side restoration after file validation; not an operator fallback.
default RmsNormResult	TensorBackend.rmsNorm(Tensor input, Tensor gamma, float epsilon)
default Tensor	TensorBackend.rmsNormBackward(Tensor gradient, Tensor normalized, Tensor rms, Tensor gamma)
static Tensor	CpuNdPosition.rotary(Tensor input, Tensor cosine, Tensor sine, int length, boolean inverse)
default Tensor	TensorBackend.rotary(Tensor input, Tensor cosine, Tensor sine, int sequenceLength, boolean inverse)
static boolean	CpuNdElementwise.sameMatrices(Tensor left, Tensor right)
static Tensor	CpuRowAccess.sample(Tensor input, int count, Random random)
static Tensor	Tensor.sampleRows(Tensor t, int n, Random rnd)
static void	Tensor.scatterAddRows(Tensor target, int[] indices, Tensor grad)
void	TensorBackend.scatterAddRows(Tensor target, int[] indices, Tensor grad)
static Tensor	CpuNdIndexing.select(Tensor input, int axis, int[] indices)
static Tensor	CpuRowAccess.select(Tensor input, int[] rows, int columns)
default Tensor	TensorBackend.select(Tensor input, int axis, int index)
void	Tensor.setRow(int row, Tensor source, int srcRow)
Tensor	TensorBackend.sigmoid(Tensor a)
void	TensorBackend.sigmoidInPlace(Tensor a)
default Tensor	TensorBackend.slice(Tensor input, int axis, int start, int stop, int step)
static Tensor	CpuNdReduction.sliceGradient(Tensor gradient, int[] dimensions, int axis, int start, int step)
default Tensor	TensorBackend.sliceGradient(Tensor gradient, int[] shape, int axis, int start, int step)
static Tensor	Tensor.sliceRows(Tensor t, int[] rowIndices, int cols)
default Tensor	TensorBackend.sliceRows(Tensor input, int[] rows)
static Tensor	CpuNdPosition.softmax(Tensor input, int length, float scale)
default Tensor	TensorBackend.softmaxAxis(Tensor input, int axis)
default Tensor	TensorBackend.softmaxAxisBackward(Tensor gradient, Tensor output, int axis)
Tensor	TensorBackend.softmaxBackward(Tensor gradOutput, Tensor softmaxOut)
default Tensor	TensorBackend.softmaxBackward(Tensor gradOutput, Tensor softmaxOut, float scale)
Tensor	TensorBackend.softmaxRows(Tensor logits)
static Tensor	HeadLayout.split(Tensor input, int heads)
static Tensor	PackedLayouts.split(TensorBackend backend, Tensor input, int heads)
default Tensor	TensorBackend.splitHeads(Tensor input, int heads)
Tensor	TensorBackend.sqrt(Tensor a)
void	TensorBackend.sqrtInPlace(Tensor a)
static void	NormalizationChecks.statistics(Tensor input, Tensor denominator)
static int[]	NormalizationChecks.statisticsShape(Tensor input)
Tensor	Tensor.subtract(Tensor other)
Tensor	TensorBackend.subtract(Tensor a, Tensor b)
Tensor	Tensor.subtractBroadcastCols(Tensor colVector)
Tensor	TensorBackend.subtractBroadcastCols(Tensor a, Tensor colVector)
Tensor	Tensor.subtractInPlace(Tensor other)
void	TensorBackend.subtractInPlace(Tensor a, Tensor b)
static Tensor	CpuNdReduction.sum(Tensor input, int[] axes, boolean keep)
float	TensorBackend.sum(Tensor a)
float	TensorBackend.sumAbs(Tensor a)
Tensor	TensorBackend.sumAlongCols(Tensor a)
Tensor	TensorBackend.sumAlongRows(Tensor a)
default Tensor	TensorBackend.sumAxes(Tensor input, int[] axes, boolean keepDimensions)
Tensor	TensorBackend.sumRows(Tensor a)
default Tensor	TensorBackend.sumToShape(Tensor input, int... shape)
boolean	CausalAttentionBackend.supportsCausalAttention(Tensor query, Tensor key, Tensor value, int heads, float scale)
default boolean	CausalAttentionBackend.supportsCausalAttention(Tensor query, Tensor key, Tensor value, int queryHeads, int keyValueHeads, float scale)
default Tensor	TensorBackend.swiGlu(Tensor gate, Tensor up)
default SwiGluBackwardResult	TensorBackend.swiGluBackward(Tensor gradient, Tensor gate, Tensor up)
Tensor	TensorBackend.tanh(Tensor a)
void	TensorBackend.tanhInPlace(Tensor a)
static float	CpuNdReduction.total(Tensor input, boolean absolute)
static int	HeadLayout.totalHeads(Tensor input, int heads)
Tensor	TensorBackend.transpose(Tensor a)
static Tensor	CpuNdElementwise.unary(Tensor input, CpuNdElementwise.Unary operation)
static void	CpuNdElementwise.unaryInto(Tensor input, Tensor output, CpuNdElementwise.Unary operation)
static void	TensorAdapters.unpackF32Into(float[] flat, Tensor t)
default Tensor	TensorBackend.unsqueeze(Tensor input, int axis)
static void	CpuNdAdamW.update(Tensor weight, Tensor gradient, Tensor first, Tensor second, float learningRate, float beta1, float beta2, float epsilon, float decay, float bc1, float bc2)
static Tensor	CpuNdStatistics.variance(Tensor input, int[] axes, boolean keep, int correction)
Tensor	TensorBackend.varianceAlongRows(Tensor a)
default Tensor	TensorBackend.varianceAxes(Tensor input, int[] axes, boolean keepDimensions, int correction)
default Tensor	TensorBackend.view(Tensor input, int... shape)
static int	NormalizationChecks.width(Tensor input)
static void	TensorEncoding.write(DataOutput output, Tensor tensor)
default void	TensorBackend.writeHeadCache(Tensor cache, Tensor update, int heads, int offset)
static void	HeadCacheOps.writeLayouts(TensorBackend backend, Tensor cache, Tensor update, int heads, int offset)
static void	TensorEncoding.writeValues(DataOutput output, Tensor tensor)

Method parameters in io.github.kirstenali.deepj.tensor with type arguments of type Tensor

Modifier and Type	Method	Description
default float	TensorBackend.l2Norm(List<Tensor> tensors)

Constructors in io.github.kirstenali.deepj.tensor with parameters of type Tensor

Modifier	Constructor	Description
	AttentionGradients(Tensor query, Tensor key, Tensor value)	Creates an instance of a AttentionGradients record class.
	CausalAttentionResult(Tensor output, Tensor normalization)
	CausalAttentionResult(Tensor output, Tensor normalization, Tensor accumulatedOutput)	Creates an instance of a CausalAttentionResult record class.
	CrossEntropyResult(Tensor loss, Tensor gradient)	Creates an instance of a CrossEntropyResult record class.
	NormalizationResult(Tensor output, Tensor normalized, Tensor denominator)	Creates an instance of a NormalizationResult record class.
	RmsNormResult(Tensor output, Tensor normalized, Tensor rms)	Creates an instance of a RmsNormResult record class.
	SwiGluBackwardResult(Tensor gateGradient, Tensor upGradient)	Creates an instance of a SwiGluBackwardResult record class.
	Tensor(Tensor source)

Uses of Tensor in io.github.kirstenali.deepj.tensor.cpu

Methods in io.github.kirstenali.deepj.tensor.cpu that return Tensor

Modifier and Type	Method	Description
Tensor	CpuBackend.batchedMatmul(Tensor left, Tensor right, int batches, boolean ta, boolean tb)
Tensor	CpuBackend.cast(Tensor input, DType dtype)
Tensor	CpuBackend.causalSoftmax(Tensor input, int length, float scale)
Tensor	CpuBackend.indexSelect(Tensor input, int axis, int[] indices)
Tensor	CpuBackend.indexSelectGradient(Tensor gradient, int[] shape, int axis, int[] indices)
Tensor	CpuBackend.maximumAxes(Tensor input, int[] axes, boolean keep)
Tensor	CpuBackend.maximumAxesBackward(Tensor input, Tensor maximum, Tensor gradient, int[] axes, boolean keep)
Tensor	CpuBackend.meanAxes(Tensor input, int[] axes, boolean keepDimensions)
Tensor	CpuBackend.mergeHeads(Tensor input, int heads)
Tensor	CpuBackend.normalizationBackward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma, boolean centered)
Tensor	CpuBackend.rmsNormBackward(Tensor gradient, Tensor normalized, Tensor rms, Tensor gamma)
Tensor	CpuBackend.sliceGradient(Tensor gradient, int[] shape, int axis, int start, int step)
Tensor	CpuBackend.sliceRows(Tensor input, int[] rows)
Tensor	CpuBackend.softmaxAxis(Tensor input, int axis)
Tensor	CpuBackend.splitHeads(Tensor input, int heads)
Tensor	CpuBackend.sumAxes(Tensor input, int[] axes, boolean keepDimensions)
Tensor	CpuBackend.swiGlu(Tensor gate, Tensor up)
Tensor	CpuBackend.varianceAxes(Tensor input, int[] axes, boolean keep, int correction)
Tensor	CpuBackend.zeros(DType dtype, int... shape)

Methods in io.github.kirstenali.deepj.tensor.cpu with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	CpuBackend.batchedMatmul(Tensor left, Tensor right, int batches, boolean ta, boolean tb)
Tensor	CpuBackend.cast(Tensor input, DType dtype)
Tensor	CpuBackend.causalSoftmax(Tensor input, int length, float scale)
void	CpuBackend.copyInto(Tensor target, Tensor source)
CrossEntropyResult	CpuBackend.crossEntropy(Tensor logits, int[] targets)
CrossEntropyResult	CpuBackend.crossEntropy(Tensor logits, int[] targets, boolean[] mask)
void	CpuBackend.indexAdd(Tensor target, int axis, int[] indices, Tensor updates)
Tensor	CpuBackend.indexSelect(Tensor input, int axis, int[] indices)
Tensor	CpuBackend.indexSelectGradient(Tensor gradient, int[] shape, int axis, int[] indices)
Tensor	CpuBackend.maximumAxes(Tensor input, int[] axes, boolean keep)
Tensor	CpuBackend.maximumAxesBackward(Tensor input, Tensor maximum, Tensor gradient, int[] axes, boolean keep)
Tensor	CpuBackend.meanAxes(Tensor input, int[] axes, boolean keepDimensions)
Tensor	CpuBackend.mergeHeads(Tensor input, int heads)
Tensor	CpuBackend.normalizationBackward(Tensor gradient, Tensor normalized, Tensor denominator, Tensor gamma, boolean centered)
NormalizationResult	CpuBackend.normalize(Tensor input, Tensor gamma, Tensor beta, float epsilon, boolean centered)
RmsNormResult	CpuBackend.rmsNorm(Tensor input, Tensor gamma, float epsilon)
Tensor	CpuBackend.rmsNormBackward(Tensor gradient, Tensor normalized, Tensor rms, Tensor gamma)
Tensor	CpuBackend.sliceGradient(Tensor gradient, int[] shape, int axis, int start, int step)
Tensor	CpuBackend.sliceRows(Tensor input, int[] rows)
Tensor	CpuBackend.softmaxAxis(Tensor input, int axis)
Tensor	CpuBackend.splitHeads(Tensor input, int heads)
Tensor	CpuBackend.sumAxes(Tensor input, int[] axes, boolean keepDimensions)
Tensor	CpuBackend.swiGlu(Tensor gate, Tensor up)
SwiGluBackwardResult	CpuBackend.swiGluBackward(Tensor gradient, Tensor gate, Tensor up)
Tensor	CpuBackend.varianceAxes(Tensor input, int[] axes, boolean keep, int correction)

Uses of Tensor in io.github.kirstenali.deepj.training

Methods in io.github.kirstenali.deepj.training with parameters of type Tensor

Modifier and Type	Method	Description
static Trainer	SupervisedTraining.trainer(Layer model, LossFunction lossFn, ParameterOptimizer opt, Tensor xAll, Tensor yAll, long seed)

Uses of Tensor in io.github.kirstenali.deepj.transformer

Methods in io.github.kirstenali.deepj.transformer that return Tensor

Modifier and Type	Method	Description
Tensor	LoopedTransformer.forward(Tensor input)
Tensor	LoopedTransformer.forward(Tensor input, InferenceContext inference)
Tensor	TransformerStack.forward(Tensor x)
Tensor	TransformerStack.forward(Tensor input, InferenceContext context)

Methods in io.github.kirstenali.deepj.transformer with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	LoopedTransformer.forward(Tensor input)
Tensor	LoopedTransformer.forward(Tensor input, InferenceContext inference)
Tensor	TransformerStack.forward(Tensor x)
Tensor	TransformerStack.forward(Tensor input, InferenceContext context)

Uses of Tensor in io.github.kirstenali.deepj.transformer.embeddings

Methods in io.github.kirstenali.deepj.transformer.embeddings that return Tensor

Modifier and Type	Method	Description
Tensor	RotaryEmbedding.apply(Tensor t, int seqLen, int nHeads)
Tensor	RotaryEmbedding.apply(Tensor tensor, int length, int heads, int offset)
Tensor	RotaryEmbedding.applyBackward(Tensor t, int seqLen, int nHeads)
Tensor	RotaryEmbedding.applyInference(Tensor input, int length, int heads, int offset)	Continuing positions for rolling inference; persistent rotary tables never grow.
Tensor	Embedding.forward(int[] ids)
Tensor	PositionalEmbedding.forward(int seqLen)
Tensor	PositionalEmbedding.forward(int seqLen, int offset)
Tensor	Embedding.forwardBatch(int[][] sequences)

Methods in io.github.kirstenali.deepj.transformer.embeddings with parameters of type Tensor

Modifier and Type	Method	Description
Tensor	RotaryEmbedding.apply(Tensor t, int seqLen, int nHeads)
Tensor	RotaryEmbedding.apply(Tensor tensor, int length, int heads, int offset)
Tensor	RotaryEmbedding.applyBackward(Tensor t, int seqLen, int nHeads)
Tensor	RotaryEmbedding.applyInference(Tensor input, int length, int heads, int offset)	Continuing positions for rolling inference; persistent rotary tables never grow.