Uses of Package
ai.beestream.jitnn.engine

Packages that use ai.beestream.jitnn.engine
Package
Description
The jitNN API: Tensor (n-dimensional arrays, lazy on TornadoVM devices), reverse-mode differentiation (Tensor.backward(), Grad), the functional operators of F, linear algebra (Linalg), Fourier transforms (Fft), sparse matrices (SparseMatrix), mixed precision (Autocast) and user-defined operators (CustomOp).
The execution engine, which records operators in a per-thread trace and, at a sync point, compiles the trace into one TornadoVM task graph, cached by structure and replayed, or runs it on the host with the double-precision reference.
  • Class
    Description
    What an operator sees while its device tasks are added to a task graph: the arrays bound to its tensors, constant metadata arrays, and helpers for launch grids.
    One operator application.
    An operator: how to run it on a device (KernelContext kernels), how to run it on the host (the reference used by tests), and how to differentiate it (in terms of other operators).
    The memory behind one or more tensors (a tensor and its views share a storage), and where its newest contents live.
  • Class
    Description
    Adds one task to the graph under the given task id.
    The libraries jitNN can call.
    One operator application.
    An operator: how to run it on a device (KernelContext kernels), how to run it on the host (the reference used by tests), and how to differentiate it (in terms of other operators).