Tensors

WaffeTensor

const

tensor

sysconst

parameter

data

value

grad

backward

Gradients

with-no-grad

*no-grad*

Four Arithmetic Operations

Element wise operations

!add Adds the given tensors element by element.

!sub Substracts the given tensors element by element.

!mul Multiply the given tensors element by element.

!div Divides the given tensors element by element.

Summarize

Get a scalar value from a matrix

!sum

!mean

Multiplying matrices

!dot

!matmul

!einsum ⚠️ It is unavailable.

Trigonometric Functions

!sin

!cos

!tan

!asin

!acos

!atan

!sinh

!cosh

!tanh

!asinh

!acosh

!atanh

Mathematical Functions

!abs

!log

!exp

!pow

!sqrt

!argmax

!argmin

Reshaping

!squeeze

!unsqueeze

!reshape

!repeats

!flatten

!transpose Often used with !matmul

!transpose1 Forced to transpose.

Shaping

Accessing the shape

!shape

!dims

!size

Concatenate and Split

!concatenate

!stack

!split

!vstack

!hstack

Iterations and Making Copy

!aref it behaves as if aref, but works like numpy

!displace TODO

!set-batch displace the tensor

!where

!index

!filter

Sampling Distributions

Initializes the tensor of the given dim with specified algorithms.

!normal Sampling the standard distribution

!randn

!uniform-random Sampling the uniform random

!beta Sampling the beta distribution

!gamma Sampling the gamma distribution

!chisquare Sampling the chisquare distribution

!bernoulli Sampling the bernoulli distribution

!binomial

Initializes the tensor

Tensors with the same elements.

!zeros

!ones

!fill Fill up a tensor with specified value

!zeros-like Returns a tensor with the same dimension as the given dimension but elements are zero.

!ones-like

!full-like

!arange

Activations

!tanh

!sigmoid

!relu

!gelu Not Tested

!leakey-relu Not Tested

!swish Not Tested

!mish TODO

!softmax

Logging

with-verbose Displays Computation Node

Dtype

with-dtype

dtypecase

define-with-typevar

Extensions

with-backend

define-node-extension

*restart-non-exist-backend*

Destructive Operations

!allow-destruct

!disallow-destruct

Objects

defmodel/defnode/defoptimizer

defnode

defmodel

defoptimizer

call

call-backward

self

save-for-backward

get-forward-caller

get-backward-caller

with-calling-layers

Trainer

deftrainer

step-model

predict

update

zero-grad

Dataset

defdataset

get-dataset

get-dataset-length

Model List

mlist

model-list

mlist

Printing

print-model

*print-char-max-size*

*print-arr-max-size*

*print-mat-max-size*

*ignore-inlining-info*