TorchLean API

Docs HomeInstallationGuideExamplesGraphsTools

NN.Runtime.Autograd.Torch.Core.Session.Random

Eager Random Tensors #

A session counter gives repeated seeded calls distinct keys, including calls after a tape reset. CPU and CUDA dispatch use the same key schedule.

source
def Runtime.Autograd.Torch.Internal.EagerSession.randUniform {α : Type} [TorchLean.Storage α] [Context α] [TensorTransfer α] (s : EagerSession α) {sh : Spec.Shape} (seed : ℕ) (name : Option String := none) :
IO (TensorRef α sh)

Record a seeded uniform tensor using the session's next random key.

Instances For
    source
    def Runtime.Autograd.Torch.Internal.EagerSession.bernoulliMask {α : Type} [TorchLean.Storage α] [Context α] [TensorTransfer α] (s : EagerSession α) {sh : Spec.Shape} (keepProb : TensorRef α Spec.Shape.scalar) (seed : ℕ) (name : Option String := none) :
    IO (TensorRef α sh)

    Deterministic {0,1} mask generator (seeded) with a scalar keep-probability input.

    Instances For