Uses of Package
io.github.kirstenali.deepj.tensor
Packages that use io.github.kirstenali.deepj.tensor
Package
Description
Neural networks composed from DeepJ tensors, layers, autograd and backend operations.
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.activations
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.autograd
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.inference
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.layers
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.layers.transformer
-
-
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.loss
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.nn
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.nn.transformer
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.nn.transformer.loop
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.nn.transformer.nova
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.nn.transformer.orbit
-
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.nn.transformer.prism
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.optimisersClassDescriptionDense floating-point storage formats.Validated logical checkpoint metadata; reading it never allocates tensor values.
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.persistenceClassDescriptionValidated logical checkpoint metadata; reading it never allocates tensor values.
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.tensorClassDescriptionInternal lifetime tracking for non-escaping buffers produced by library backward rules.Output, row statistics and unrounded FP32 output needed by tiled backward.Dense floating-point storage formats.The output uses promoted storage; normalized values and denominators retain FLOAT32 accumulation.Logical dimensions for the compatibility API; never inferred from a GPU allocation's physical rows.Validated logical checkpoint metadata; reading it never allocates tensor values.Shape transformations retain one Tensor system and participate in first-order autograd.Bounds internal view lifetimes, without releasing saved or explicitly retained storage.Shared physical storage plumbing; logical Tensor operations remain on the concrete facade.
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.tensor.cpuClassDescriptionDense floating-point storage formats.The output uses promoted storage; normalized values and denominators retain FLOAT32 accumulation.
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.tensor.metalClassDescriptionOptional backend capability for memory-efficient, fused causal attention.
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.training
-
Classes in io.github.kirstenali.deepj.tensor used by io.github.kirstenali.deepj.transformer
-