Sandbox::https://github.com/charlescearl/DeepRacket/blob/master/tensor.rkt

Repository navigation

;;Typed tensors

#lang typed/racket

(require/typed

"lib-cudann.rkt"

ffi/unsafe

math/base

math/matrix)

;; define Tensor struct

;; Provide an array, then store the FlVector of it and pointer

;; Create a gpu vector of size for the representation

;; We then memcpy between this vector and gpu vector to return

;; values.

;; desc stores the tensordescriptor

(struct tensor ([in-vect : FlVector]

[src-ptr : CPointer]

[gpu-vect : CPointer]

[desc : CType]))

;; Tensor

;; [ [ [1 2 3] [3 4 5] [3 4 5] [3 4 5] ] [ [5 6 7] [5 6 7] [5 6 7] [7 8 9]]]

;; batch size is 2, sequence length is 4

;; input size is 3

;; For input layers

(: make-input-tensor (-> Integer tensor))

(define (make-input-tensor array-size)

(let ([ptr (init-ptr)])

(cudaMalloc ptr size)

(tensor ptr)))

;; For hidden layers of the network

(: make-layer-tensor (-> Integer tensor))

(define (make-layer-tensor array-size)

(let ([ptr (init-ptr)])

(cudaMalloc ptr size)

(tensor ptr)))