Description
R Bindings to 'PyTorch'.
Description
'R' implementation and interface of the Machine Learning platform 'PyTorch' <https://pytorch.org/> developed in 'Python'. It requires a 'conda' environment with 'torch' and 'torchvision' Python packages to provide 'PyTorch' functions, methods and classes. The key object in 'PyTorch' is the tensor which is in essence a multidimensional array. These tensors are fairly flexible in performing calculations in CPUs as well as 'GPUs' to accelerate tensor operations.
README.md
rTorch
The goal of rTorch is providing an R wrapper to PyTorch. rTorch provides all the functionality of PyTorch plus all the features that R provides. We have borrowed some ideas and code used in R tensorflow to implement rTorch.
Besides the module torch, which directly provides PyTorch methods, classes and functions, the package also provides the modules numpy as a method called np, and torchvision, as well. The dollar sign $ after the module will provide you access to all their sub-objects. Example:
tv <- rTorch::torchvision
tv
#> Module(torchvision)
np <- rTorch::np
np
#> Module(numpy)
torch_module <- rTorch::torch
torch_module
#> Module(torch)
rTorch Examples
To lighten up the time in building this rTorch package, we moved the examples that use tensor operations and neural networks to separate repositories. There are two sets of examples:
- The rTorch Minimal Book at https://f0nzie.github.io/rtorch-minimal-book/
- rTorch Advanced Examples at https://github.com/f0nzie/rTorch.examples
rTorch installation
rTorch is available via CRAN and from GitHub.
From CRAN
Install from CRAN using install.packages("rTorch") from the R console, or from RStudio using Tools, Install Packages from the menu.
From GitHub
For the latest version install from GitHub. Install rTorch with:
devtools::install_github("f0nzie/rTorch")
Installing from GitHub gives you the flexibility of experimenting with the latest development version of rTorch. For instance, to install rTorch from the develop branch:
devtools::install_github("f0nzie/rTorch", ref="develop")
or clone with Git with:
git clone https://github.com/f0nzie/rTorch.git
Getting Started
Tensor types
There are five major type of Tensors in PyTorch: * Byte * Float * Double * Long * Bool
library(rTorch)
byte_tensor <- torch$ByteTensor(3L, 3L)
float_tensor <- torch$FloatTensor(3L, 3L)
double_tensor <- torch$DoubleTensor(3L, 3L)
long_tensor <- torch$LongTensor(3L, 3L)
bool_tensor <- torch$BoolTensor(5L, 5L)
byte_tensor
#> tensor([[58, 0, 0],
#> [ 0, 0, 0],
#> [ 0, 0, 0]], dtype=torch.uint8)
float_tensor
#> tensor([[0., 0., 0.],
#> [0., 0., 0.],
#> [0., 0., 0.]])
double_tensor
#> tensor([[6.9259e-310, 6.9259e-310, 1.2095e-312],
#> [1.2697e-321, 6.9259e-310, 4.6888e-310],
#> [ 0.0000e+00, 0.0000e+00, 0.0000e+00]], dtype=torch.float64)
long_tensor
#> tensor([[140182629985840, 140182629985840, 0],
#> [ 94902887985680, 0, 0],
#> [ 0, 0, 0]])
bool_tensor
#> tensor([[False, False, False, False, False],
#> [False, False, False, False, False],
#> [False, False, False, False, False],
#> [False, False, False, False, False],
#> [False, False, False, False, False]])
A 4D tensor like in MNIST hand-written digits recognition dataset:
mnist_4d <- torch$FloatTensor(60000L, 3L, 28L, 28L)
# size
mnist_4d$size()
#> torch.Size([60000, 3, 28, 28])
# length
length(mnist_4d)
#> [1] 141120000
# shape, like in numpy
mnist_4d$shape
#> torch.Size([60000, 3, 28, 28])
# number of elements
mnist_4d$numel()
#> [1] 141120000
A 3D tensor:
ft3d <- torch$FloatTensor(4L, 3L, 2L)
ft3d
#> tensor([[[0., 0.],
#> [0., 0.],
#> [0., 0.]],
#>
#> [[0., 0.],
#> [0., 0.],
#> [0., 0.]],
#>
#> [[0., 0.],
#> [0., 0.],
#> [0., 0.]],
#>
#> [[0., 0.],
#> [0., 0.],
#> [0., 0.]]])
# get first element in a tensor
ft3d[1, 1, 1]
#> tensor(0.)
# create a tensor with a value
torch$full(list(2L, 3L), 3.141592)
#> tensor([[3.1416, 3.1416, 3.1416],
#> [3.1416, 3.1416, 3.1416]])
Basic Tensor Operations
Add tensors
# 3x5 matrix uniformly distributed between 0 and 1
mat0 <- torch$FloatTensor(3L, 5L)$uniform_(0L, 1L)
# fill a 3x5 matrix with 0.1
mat1 <- torch$FloatTensor(3L, 5L)$uniform_(0.1, 0.1)
# a vector with all ones
mat2 <- torch$FloatTensor(5L)$uniform_(1, 1)
# add two tensors
mat0 + mat1
#> tensor([[1.0472, 0.3601, 0.8105, 0.2573, 0.9311],
#> [0.8394, 0.1631, 0.5836, 0.2908, 0.1806],
#> [1.0795, 0.7679, 0.7024, 0.1542, 1.0788]])
# add three tensors
mat0 + mat1 + mat2
#> tensor([[2.0472, 1.3601, 1.8105, 1.2573, 1.9311],
#> [1.8394, 1.1631, 1.5836, 1.2908, 1.1806],
#> [2.0795, 1.7679, 1.7024, 1.1542, 2.0788]])
# PyTorch add two tensors using add() function
x = torch$rand(5L, 4L)
y = torch$rand(5L, 4L)
print(x$add(y))
#> tensor([[1.3989, 1.0835, 0.5394, 1.1483],
#> [1.3036, 0.3943, 1.2853, 1.2172],
#> [1.0630, 1.6868, 1.1131, 1.5882],
#> [0.5632, 0.6763, 1.1735, 1.5093],
#> [1.5244, 0.9094, 1.4205, 0.5644]])
print(x + y)
#> tensor([[1.3989, 1.0835, 0.5394, 1.1483],
#> [1.3036, 0.3943, 1.2853, 1.2172],
#> [1.0630, 1.6868, 1.1131, 1.5882],
#> [0.5632, 0.6763, 1.1735, 1.5093],
#> [1.5244, 0.9094, 1.4205, 0.5644]])
Add a tensor element to another tensor
# add an element of a tensor to another tensor
mat1[1, 1] + mat2
#> tensor([1.1000, 1.1000, 1.1000, 1.1000, 1.1000])
mat1
#> tensor([[0.1000, 0.1000, 0.1000, 0.1000, 0.1000],
#> [0.1000, 0.1000, 0.1000, 0.1000, 0.1000],
#> [0.1000, 0.1000, 0.1000, 0.1000, 0.1000]])
# extract part of the tensor
indices <- torch$tensor(c(0L, 3L))
torch$index_select(mat1, 1L, indices) # rows = 0; columns = 1
#> tensor([[0.1000, 0.1000],
#> [0.1000, 0.1000],
#> [0.1000, 0.1000]])
Add a scalar to a tensor
# add a scalar to a tensor
mat0 + 0.1
#> tensor([[1.0472, 0.3601, 0.8105, 0.2573, 0.9311],
#> [0.8394, 0.1631, 0.5836, 0.2908, 0.1806],
#> [1.0795, 0.7679, 0.7024, 0.1542, 1.0788]])
Multiply a tensor by a scalar
# Multiply tensor by scalar
tensor = torch$ones(4L, dtype=torch$float64)
scalar = np$float64(4.321)
message("a numpy scalar: ", scalar)
#> a numpy scalar: 4.321
message("a PyTorch scalar: ", torch$scalar_tensor(scalar))
#> a PyTorch scalar: tensor(4.3210)
message("\nResult")
#>
#> Result
(prod = torch$mul(tensor, torch$scalar_tensor(scalar)))
#> tensor([4.3210, 4.3210, 4.3210, 4.3210], dtype=torch.float64)
# short version using generics
(prod = tensor * scalar)
#> tensor([4.3210, 4.3210, 4.3210, 4.3210], dtype=torch.float64)
Multiply two 1D tensors
t1 = torch$tensor(c(1, 2))
t2 = torch$tensor(c(3, 2))
t1
#> tensor([1., 2.])
t2
#> tensor([3., 2.])
t1 * t2
#> tensor([3., 4.])
t1 = torch$tensor(list(
c(1, 2, 3),
c(1, 2, 3)
))
t2 = torch$tensor(list(
c(1, 2),
c(1, 2),
c(1, 2)
))
t1
#> tensor([[1., 2., 3.],
#> [1., 2., 3.]])
t2
#> tensor([[1., 2.],
#> [1., 2.],
#> [1., 2.]])
torch$mm(t1, t2)
#> tensor([[ 6., 12.],
#> [ 6., 12.]])
Dot product for 1D tensors (vectors)
t1 = torch$tensor(c(1, 2))
t2 = torch$tensor(c(3, 2))
t1
#> tensor([1., 2.])
t2
#> tensor([3., 2.])
# dot product of two vectors
torch$dot(t1, t2)
#> tensor(7.)
# Dot product of 1D tensors is a scalar
p <- torch$Tensor(list(4L, 2L))
q <- torch$Tensor(list(3L, 1L))
(r = torch$dot(p, q)) # 14
#> tensor(14.)
(r <- p %.*% q)
#> tensor(14.)
# torch$dot product will work for vectors not matrices
t1 = torch$tensor(list(
c(1, 2, 3),
c(1, 2, 3)
))
t2 = torch$tensor(list(
c(1, 2),
c(1, 2),
c(1, 2)
))
t1$shape
#> torch.Size([2, 3])
t2$shape
#> torch.Size([3, 2])
# RuntimeError: 1D tensors expected, got 2D, 2D tensors
torch$dot(t1, t2)
Dot product for 2D tensors (matrices)
The number of columns of the first matrix must be equal to the number of rows of the second matrix.
# for the dot product of nD tensors we use torch$mm()
t1 = torch$tensor(list(
c(1, 2, 3),
c(1, 2, 3)
))
t2 = torch$tensor(list(
c(1, 2),
c(1, 2),
c(1, 2)
))
torch$mm(t1, t2)
#> tensor([[ 6., 12.],
#> [ 6., 12.]])
torch$mm(t2, t1)
#> tensor([[3., 6., 9.],
#> [3., 6., 9.],
#> [3., 6., 9.]])
# for the dot product of 2D tensors we use torch$mm()
t1 = torch$arange(1, 11)$view(c(2L,5L))
t2 = torch$arange(11, 21)$view(c(5L,2L))
t1
#> tensor([[ 1., 2., 3., 4., 5.],
#> [ 6., 7., 8., 9., 10.]])
t2
#> tensor([[11., 12.],
#> [13., 14.],
#> [15., 16.],
#> [17., 18.],
#> [19., 20.]])
# result
torch$mm(t1, t2)
#> tensor([[245., 260.],
#> [620., 660.]])
Multiplication for nD tensors
# 1D tensor
t1 = torch$tensor(c(1, 2))
t2 = torch$tensor(c(3, 2))
torch$matmul(t1, t2)
#> tensor(7.)
# 2D tensor
t1 = torch$tensor(list(
c(1, 2, 3),
c(1, 2, 3)
))
t2 = torch$tensor(list(
c(1, 2),
c(1, 2),
c(1, 2)
))
torch$matmul(t1, t2)
#> tensor([[ 6., 12.],
#> [ 6., 12.]])
# for the dot product of 3D tensors we use torch$matmul()
t1 = torch$arange(1, 13)$view(c(2L, 2L, 3L)) # number of columns = 2
t2 = torch$arange(0, 18)$view(c(2L, 3L, 3L)) # number of rows = 2
t1
#> tensor([[[ 1., 2., 3.],
#> [ 4., 5., 6.]],
#>
#> [[ 7., 8., 9.],
#> [10., 11., 12.]]])
t2
#> tensor([[[ 0., 1., 2.],
#> [ 3., 4., 5.],
#> [ 6., 7., 8.]],
#>
#> [[ 9., 10., 11.],
#> [12., 13., 14.],
#> [15., 16., 17.]]])
message("result")
#> result
torch$matmul(t1, t2)
#> tensor([[[ 24., 30., 36.],
#> [ 51., 66., 81.]],
#>
#> [[294., 318., 342.],
#> [402., 435., 468.]]])
t1 = torch$arange(1, 13)$view(c(3L, 2L, 2L)) # number of columns = 3
t2 = torch$arange(0, 12)$view(c(3L, 2L, 2L)) # number of rows = 3
t1
#> tensor([[[ 1., 2.],
#> [ 3., 4.]],
#>
#> [[ 5., 6.],
#> [ 7., 8.]],
#>
#> [[ 9., 10.],
#> [11., 12.]]])
t2
#> tensor([[[ 0., 1.],
#> [ 2., 3.]],
#>
#> [[ 4., 5.],
#> [ 6., 7.]],
#>
#> [[ 8., 9.],
#> [10., 11.]]])
message("result")
#> result
torch$matmul(t1, t2)
#> tensor([[[ 4., 7.],
#> [ 8., 15.]],
#>
#> [[ 56., 67.],
#> [ 76., 91.]],
#>
#> [[172., 191.],
#> [208., 231.]]])
cross product
m1 = torch$ones(3L, 5L)
m2 = torch$ones(3L, 5L)
v1 = torch$ones(3L)
# Cross product
# Size 3x5
(r = torch$cross(m1, m2))
#> tensor([[0., 0., 0., 0., 0.],
#> [0., 0., 0., 0., 0.],
#> [0., 0., 0., 0., 0.]])
NumPy and PyTorch
numpy has been made available as a module inside rTorch. We could call functions from numpy refrerring to it as np$any_function. Examples:
# a 2D numpy array
syn0 <- np$random$rand(3L, 5L)
syn0
#> [,1] [,2] [,3] [,4] [,5]
#> [1,] 0.1524218 0.7769113 0.6153864 0.00220404 0.78412198
#> [2,] 0.4959399 0.7230621 0.9840282 0.64843544 0.06556167
#> [3,] 0.5931231 0.6513373 0.4399219 0.57722973 0.94843503
# numpy arrays of zeros
syn1 <- np$zeros(c(5L, 10L))
syn1
#> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
#> [1,] 0 0 0 0 0 0 0 0 0 0
#> [2,] 0 0 0 0 0 0 0 0 0 0
#> [3,] 0 0 0 0 0 0 0 0 0 0
#> [4,] 0 0 0 0 0 0 0 0 0 0
#> [5,] 0 0 0 0 0 0 0 0 0 0
# add a scalar to a numpy array
syn1 = syn1 + 0.1
syn1
#> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
#> [1,] 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
#> [2,] 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
#> [3,] 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
#> [4,] 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
#> [5,] 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
# in numpy a multidimensional array needs to be defined with a tuple
# From R we use a vector to refer to a tuple in Python
l1 <- np$ones(c(5L, 5L))
l1
#> [,1] [,2] [,3] [,4] [,5]
#> [1,] 1 1 1 1 1
#> [2,] 1 1 1 1 1
#> [3,] 1 1 1 1 1
#> [4,] 1 1 1 1 1
#> [5,] 1 1 1 1 1
# vector-matrix multiplication
np$dot(syn0, syn1)
#> [,1] [,2] [,3] [,4] [,5] [,6] [,7]
#> [1,] 0.2331045 0.2331045 0.2331045 0.2331045 0.2331045 0.2331045 0.2331045
#> [2,] 0.2917027 0.2917027 0.2917027 0.2917027 0.2917027 0.2917027 0.2917027
#> [3,] 0.3210047 0.3210047 0.3210047 0.3210047 0.3210047 0.3210047 0.3210047
#> [,8] [,9] [,10]
#> [1,] 0.2331045 0.2331045 0.2331045
#> [2,] 0.2917027 0.2917027 0.2917027
#> [3,] 0.3210047 0.3210047 0.3210047
# build a numpy array from three R vectors
X <- np$array(rbind(c(1,2,3), c(4,5,6), c(7,8,9)))
X
#> [,1] [,2] [,3]
#> [1,] 1 2 3
#> [2,] 4 5 6
#> [3,] 7 8 9
# transpose the array
np$transpose(X)
#> [,1] [,2] [,3]
#> [1,] 1 4 7
#> [2,] 2 5 8
#> [3,] 3 6 9
Copying a numpy object
With newer PyTorch versions we should work with NumPy array copies There have been minor changes in the latest versions of PyTorch that prevents a direct use of a NumPy array. You will get this warning:
sys:1: UserWarning: The given NumPy array is not writeable, and PyTorch does
not support non-writeable tensors. This means you can write to the underlying
(supposedly non-writeable) NumPy array using the tensor. You may want to copy
the array to protect its data or make it writeable before converting it to a
tensor. This type of warning will be suppressed for the rest of this program.
For instance, this code will produce the warning:
# as_tensor. Modifying tensor modifies numpy object as well
a = np$array(list(1, 2, 3))
t = torch$as_tensor(a)
print(t)
torch$tensor(list( 1, 2, 3))
t[1L]$fill_(-1)
print(a)
while this other one -with some extra code- will not:
a = np$array(list(1, 2, 3))
a_copy = r_to_py(a)$copy() # we make a copy of the numpy array first
t = torch$as_tensor(a_copy)
print(t)
#> tensor([1., 2., 3.], dtype=torch.float64)
torch$tensor(list( 1, 2, 3))
#> tensor([1., 2., 3.])
t[1L]$fill_(-1)
#> tensor(-1., dtype=torch.float64)
print(a)
#> [1] 1 2 3
Function make_copy()
To make easier to copy an object in rTorch we implemented the function make_copy, which makes a safe copy regardless if it is a torch, numpy or an R type object.
a = np$array(list(1, 2, 3, 4, 5))
a_copy <- make_copy(a)
t <- torch$as_tensor(a_copy)
t
#> tensor([1., 2., 3., 4., 5.], dtype=torch.float64)
Convert a numpy array to a tensor
# convert a numpy array to a tensor
np_a = np$array(c(c(3, 4), c(3, 6)))
t_a = torch$from_numpy(r_to_py(np_a)$copy())
print(t_a)
#> tensor([3., 4., 3., 6.], dtype=torch.float64)
Creating tensors
Random tensor
# a random 1D tensor
np_arr <- np$random$rand(5L)
ft1 <- torch$FloatTensor(r_to_py(np_arr)$copy()) # make a copy of numpy array
ft1
#> tensor([0.9408, 0.8752, 0.5924, 0.7329, 0.6719])
# tensor as a float of 64-bits
np_copy <- r_to_py(np$random$rand(5L))$copy() # make a copy of numpy array
ft2 <- torch$as_tensor(np_copy, dtype= torch$float64)
ft2
#> tensor([0.0462, 0.5851, 0.4886, 0.0725, 0.8959], dtype=torch.float64)
This is a very common operation in machine learning:
# convert tensor to a numpy array
a = torch$rand(5L, 4L)
b = a$numpy()
print(b)
#> [,1] [,2] [,3] [,4]
#> [1,] 0.8167720 0.8075168 0.2668687105 0.7421414
#> [2,] 0.6829966 0.5185235 0.0005332828 0.9414444
#> [3,] 0.2030416 0.7496545 0.0358363986 0.3475423
#> [4,] 0.7263991 0.6163300 0.2169904113 0.9363614
#> [5,] 0.4336911 0.5996053 0.2127178907 0.8461853
Change the type of a tensor
# convert tensor to float 16-bits
ft2_dbl <- torch$as_tensor(ft2, dtype = torch$float16)
ft2_dbl
#> tensor([0.0462, 0.5850, 0.4885, 0.0724, 0.8960], dtype=torch.float16)
Create an uninitialized tensor
Create a tensor of size (5 x 7) with uninitialized memory:
a <- torch$FloatTensor(5L, 7L)
print(a)
#> tensor([[1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45,
#> 1.4013e-45],
#> [1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45,
#> 1.4013e-45],
#> [1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45,
#> 1.4013e-45],
#> [1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45, 1.4013e-45,
#> 1.4013e-45],
#> [1.4013e-45, 1.4013e-45, 0.0000e+00, 1.4013e-45, 1.4013e-45, 1.4013e-45,
#> 1.4013e-45]])
Create a tensor and then change its shape
# using arange to create tensor. starts from 0
v = torch$arange(9L)
(v = v$view(3L, 3L))
#> tensor([[0, 1, 2],
#> [3, 4, 5],
#> [6, 7, 8]])
Distributions
Initialize a tensor randomized with a normal distribution with mean=0, var=1:
a <- torch$randn(5L, 7L)
print(a)
#> tensor([[-0.9889, 0.6058, 0.5492, -0.0291, 1.3752, -0.2528, -1.8089],
#> [ 0.6621, -0.4370, 1.5590, 0.8149, 0.4004, -0.5299, 0.9275],
#> [ 0.0880, 0.5931, 0.2601, 1.6336, 1.0154, 1.2189, -1.6278],
#> [ 1.2171, 0.1377, -0.2377, 0.0792, -0.2885, 0.6316, 1.7481],
#> [-0.7538, 0.6162, 1.3023, -1.5574, 0.1196, -1.1652, 1.5082]])
print(a$size())
#> torch.Size([5, 7])
Uniform matrix
library(rTorch)
# 3x5 matrix uniformly distributed between 0 and 1
mat0 <- torch$FloatTensor(3L, 5L)$uniform_(0L, 1L)
# fill a 3x5 matrix with 0.1
mat1 <- torch$FloatTensor(3L, 5L)$uniform_(0.1, 0.1)
# a vector with all ones
mat2 <- torch$FloatTensor(5L)$uniform_(1, 1)
mat0
#> tensor([[0.9587, 0.3368, 0.3534, 0.0709, 0.6827],
#> [0.5521, 0.6874, 0.6756, 0.3705, 0.9120],
#> [0.4415, 0.9895, 0.4699, 0.5890, 0.0901]])
mat1
#> tensor([[0.1000, 0.1000, 0.1000, 0.1000, 0.1000],
#> [0.1000, 0.1000, 0.1000, 0.1000, 0.1000],
#> [0.1000, 0.1000, 0.1000, 0.1000, 0.1000]])
Binomial distribution
Binomial <- torch$distributions$binomial$Binomial
m = Binomial(100, torch$tensor(list(0 , .2, .8, 1)))
(x = m$sample())
#> tensor([ 0., 29., 80., 100.])
m = Binomial(torch$tensor(list(list(5.), list(10.))),
torch$tensor(list(0.5, 0.8)))
(x = m$sample())
#> tensor([[0., 4.],
#> [2., 8.]])
Exponential distribution
Exponential <- torch$distributions$exponential$Exponential
m = Exponential(torch$tensor(list(1.0)))
m$sample() # Exponential distributed with rate=1
#> tensor([0.4214])
Weibull distribution
Weibull <- torch$distributions$weibull$Weibull
m = Weibull(torch$tensor(list(1.0)), torch$tensor(list(1.0)))
m$sample() # sample from a Weibull distribution with scale=1, concentration=1
#> tensor([0.0745])
Tensor default data types
Only floating-point types are supported as the default type.
float32
# Default data type
torch$tensor(list(1.2, 3))$dtype # default for floating point is torch.float32
#> torch.float32
float64
# change default data type to float64
torch$set_default_dtype(torch$float64)
torch$tensor(list(1.2, 3))$dtype # a new floating point tensor
#> torch.float64
double
torch$set_default_dtype(torch$double)
torch$tensor(list(1.2, 3))$dtype
#> torch.float64
Tensor resizing
Using view
x = torch$randn(2L, 3L) # Size 2x3
y = x$view(6L) # Resize x to size 6
z = x$view(-1L, 2L) # Size 3x2
print(y)
#> tensor([ 0.8073, -0.7656, 1.0641, 0.3801, 0.5983, 0.7950])
print(z)
#> tensor([[ 0.8073, -0.7656],
#> [ 1.0641, 0.3801],
#> [ 0.5983, 0.7950]])
# 0 1 2
# 3 4 5
# 6 7 8
v = torch$arange(9L)
(v = v$view(3L, 3L))
#> tensor([[0, 1, 2],
#> [3, 4, 5],
#> [6, 7, 8]])
Concatenating tensors
# concatenate tensors
x = torch$randn(2L, 3L)
print(x)
#> tensor([[-0.6563, 1.5943, -0.0617],
#> [ 0.5502, 1.6150, -2.0000]])
# concatenate tensors by dim=0"
torch$cat(list(x, x, x), 0L)
#> tensor([[-0.6563, 1.5943, -0.0617],
#> [ 0.5502, 1.6150, -2.0000],
#> [-0.6563, 1.5943, -0.0617],
#> [ 0.5502, 1.6150, -2.0000],
#> [-0.6563, 1.5943, -0.0617],
#> [ 0.5502, 1.6150, -2.0000]])
# concatenate tensors by dim=1
torch$cat(list(x, x, x), 1L)
#> tensor([[-0.6563, 1.5943, -0.0617, -0.6563, 1.5943, -0.0617, -0.6563, 1.5943,
#> -0.0617],
#> [ 0.5502, 1.6150, -2.0000, 0.5502, 1.6150, -2.0000, 0.5502, 1.6150,
#> -2.0000]])
Reshape tensors
# ----- Reshape tensors -----
img <- torch$ones(3L, 28L, 28L)
print(img$size())
#> torch.Size([3, 28, 28])
img_chunks <- torch$chunk(img, chunks = 3L, dim = 0L)
print(length(img_chunks))
#> [1] 3
# 1st chunk member
img_chunk_1 <- img_chunks[[1]]
print(img_chunk_1$size())
#> torch.Size([1, 28, 28])
print(img_chunk_1$sum())
#> tensor(784.)
# 2nd chunk member
img_chunk_1 <- img_chunks[[2]]
print(img_chunk_1$size())
#> torch.Size([1, 28, 28])
print(img_chunk_1$sum())
#> tensor(784.)
# index_select. get layer 1
indices = torch$tensor(c(0L))
img2 <- torch$index_select(img, dim = 0L, index = indices)
print(img2$size())
#> torch.Size([1, 28, 28])
print(img2$sum())
#> tensor(784.)
# index_select. get layer 2
indices = torch$tensor(c(1L))
img2 <- torch$index_select(img, dim = 0L, index = indices)
print(img2$size())
#> torch.Size([1, 28, 28])
print(img2$sum())
#> tensor(784.)
# index_select. get layer 3
indices = torch$tensor(c(2L))
img2 <- torch$index_select(img, dim = 0L, index = indices)
print(img2$size())
#> torch.Size([1, 28, 28])
print(img2$sum())
#> tensor(784.)
Special tensors
Identity matrix
# identity matrix
eye = torch$eye(3L) # Create an identity 3x3 tensor
print(eye)
#> tensor([[1., 0., 0.],
#> [0., 1., 0.],
#> [0., 0., 1.]])
Ones
(v = torch$ones(10L)) # A tensor of size 10 containing all ones
#> tensor([1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])
(v = torch$ones(2L, 1L, 2L, 1L)) # Size 2x1x2x1
#> tensor([[[[1.],
#> [1.]]],
#>
#>
#> [[[1.],
#> [1.]]]])
Eye
v = torch$ones_like(eye) # A tensor with same shape as eye. Fill it with 1.
v
#> tensor([[1., 1., 1.],
#> [1., 1., 1.],
#> [1., 1., 1.]])
Zeros
(z = torch$zeros(10L)) # A tensor of size 10 containing all zeros
#> tensor([0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])
Tensor fill
Fill with a unique value
# a tensor filled with ones
(v = torch$ones(3L, 3L))
#> tensor([[1., 1., 1.],
#> [1., 1., 1.],
#> [1., 1., 1.]])
Change the tensor values by rows
# change two rows in the tensor
# we are using 1-based index
v[2L, ]$fill_(2L) # fill row 1 with 2s
#> tensor([2., 2., 2.])
v[3L, ]$fill_(3L) # fill row 2 with 3s
#> tensor([3., 3., 3.])
print(v)
#> tensor([[1., 1., 1.],
#> [2., 2., 2.],
#> [3., 3., 3.]])
Fill a tensor with a set increment
# Initialize Tensor with a range of values
(v = torch$arange(10L)) # similar to range(5) but creating a Tensor
#> tensor([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
(v = torch$arange(0L, 10L, step = 1L)) # Size 5. Similar to range(0, 5, 1)
#> tensor([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
With decimal increments
u <- torch$arange(0, 10, step = 0.5)
u
#> tensor([0.0000, 0.5000, 1.0000, 1.5000, 2.0000, 2.5000, 3.0000, 3.5000, 4.0000,
#> 4.5000, 5.0000, 5.5000, 6.0000, 6.5000, 7.0000, 7.5000, 8.0000, 8.5000,
#> 9.0000, 9.5000])
Including the ending value
# range of values with increments including the end value
start <- 0
end <- 10
step <- 0.25
w <- torch$arange(start, end+step, step)
w
#> tensor([ 0.0000, 0.2500, 0.5000, 0.7500, 1.0000, 1.2500, 1.5000, 1.7500,
#> 2.0000, 2.2500, 2.5000, 2.7500, 3.0000, 3.2500, 3.5000, 3.7500,
#> 4.0000, 4.2500, 4.5000, 4.7500, 5.0000, 5.2500, 5.5000, 5.7500,
#> 6.0000, 6.2500, 6.5000, 6.7500, 7.0000, 7.2500, 7.5000, 7.7500,
#> 8.0000, 8.2500, 8.5000, 8.7500, 9.0000, 9.2500, 9.5000, 9.7500,
#> 10.0000])
Initialize a linear or log scale Tensor
# Initialize a linear or log scale Tensor
# Create a Tensor with 10 linear points for (1, 10) inclusively
(v = torch$linspace(1L, 10L, steps = 10L))
#> tensor([ 1., 2., 3., 4., 5., 6., 7., 8., 9., 10.])
# Size 5: 1.0e-10 1.0e-05 1.0e+00, 1.0e+05, 1.0e+10
(v = torch$logspace(start=-10L, end = 10L, steps = 5L))
#> tensor([1.0000e-10, 1.0000e-05, 1.0000e+00, 1.0000e+05, 1.0000e+10])
In-place / Not-in-place
a = torch$rand(5L, 4L)
print(class(a))
#> [1] "torch.Tensor" "torch._C._TensorBase" "python.builtin.object"
# converting the tensor to a numpy array, R automatically converts it
b = a$numpy()
print(class(b))
#> [1] "matrix"
a$fill_(3.5)
#> tensor([[3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000]])
# a has now been filled with the value 3.5
# add a scalar to a tensor.
# notice that was auto-converted from an array to a tensor
b <- a$add(4.0)
# a is still filled with 3.5
# new tensor b is returned with values 3.5 + 4.0 = 7.5
print(a)
#> tensor([[3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000]])
print(b)
#> tensor([[7.5000, 7.5000, 7.5000, 7.5000],
#> [7.5000, 7.5000, 7.5000, 7.5000],
#> [7.5000, 7.5000, 7.5000, 7.5000],
#> [7.5000, 7.5000, 7.5000, 7.5000],
#> [7.5000, 7.5000, 7.5000, 7.5000]])
Tensor element assigment not implemented yet
# this will throw an error because we don't still have a function for assignment
a[1, 1] <- 7.7
print(a)
# Error in a[1, 1] <- 7.7 : object of type 'environment' is not subsettable
# This would be the right wayy to assign a value to a tensor element
a[1, 1]$fill_(7.7)
#> tensor(7.7000)
# we can see that the first element has been changed
a
#> tensor([[7.7000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000],
#> [3.5000, 3.5000, 3.5000, 3.5000]])
Some operations like
narrowdo not have in-place versions, and hence,.narrow_does not exist. Similarly, some operations likefill_do not have an out-of-place version, so.filldoes not exist.
# a[[0L, 3L]]
a[1, 4]
#> tensor(3.5000)
Access to tensor elements
Change a tensor element given its index
# replace an element at position 0, 0
(new_tensor = torch$Tensor(list(list(1, 2), list(3, 4))))
#> tensor([[1., 2.],
#> [3., 4.]])
# first row, firt column
print(new_tensor[1L, 1L])
#> tensor(1.)
# change row 1, col 1 with value of 5
new_tensor[1L, 1L]$fill_(5)
#> tensor(5.)
# which is the same as doing this
new_tensor[1, 1]$fill_(5)
#> tensor(5.)
Notice that the element was changed in-place because of
fill_.
In R the index is 1-based
print(new_tensor) # tensor([[ 5., 2.],[ 3., 4.]])
#> tensor([[5., 2.],
#> [3., 4.]])
# access an element at position (1, 0), 0-based index
print(new_tensor[2L, 1L]) # tensor([ 3.])
#> tensor(3.)
# convert it to a scalar value
print(new_tensor[2L, 1L]$item()) # 3.
#> [1] 3
# which is the same as
print(new_tensor[2, 1])
#> tensor(3.)
# and the scalar
print(new_tensor[2, 1]$item())
#> [1] 3
Extract part of a tensor
# Select indices
x = torch$randn(3L, 4L)
print(x)
#> tensor([[-2.4324, 0.5563, 1.3308, -0.6363],
#> [ 1.1925, 2.4744, -0.0463, -2.2281],
#> [-1.5476, 1.1377, 0.3645, -0.8908]])
# extract first and third row
# Select indices, dim=0
indices = torch$tensor(list(0L, 2L))
torch$index_select(x, 0L, indices)
#> tensor([[-2.4324, 0.5563, 1.3308, -0.6363],
#> [-1.5476, 1.1377, 0.3645, -0.8908]])
# extract first and third column
# Select indices, dim=1
torch$index_select(x, 1L, indices)
#> tensor([[-2.4324, 1.3308],
#> [ 1.1925, -0.0463],
#> [-1.5476, 0.3645]])
# Take by indices
src = torch$tensor(list(list(4, 3, 5),
list(6, 7, 8)) )
print(src)
#> tensor([[4., 3., 5.],
#> [6., 7., 8.]])
print( torch$take(src, torch$tensor(list(0L, 2L, 5L))) )
#> tensor([4., 5., 8.])
Tensor operations
Transpose
# two dimensions: 3x3
x <- torch$arange(9L)
x <- x$view(c(3L, 3L))
t <- torch$transpose(x, 0L, 1L)
x # "Original tensor"
#> tensor([[0, 1, 2],
#> [3, 4, 5],
#> [6, 7, 8]])
t # "Transposed"
#> tensor([[0, 3, 6],
#> [1, 4, 7],
#> [2, 5, 8]])
# three dimensions: 1x2x3
x <- torch$ones(c(1L, 2L, 3L))
t <- torch$transpose(x, 1L, 0L)
print(x) # original tensor
#> tensor([[[1., 1., 1.],
#> [1., 1., 1.]]])
print(t) # transposed
#> tensor([[[1., 1., 1.]],
#>
#> [[1., 1., 1.]]])
print(x$shape) # original tensor
#> torch.Size([1, 2, 3])
print(t$shape) # transposed
#> torch.Size([2, 1, 3])
Permutation
permute a 2D tensor
x <- torch$tensor(list(list(list(1,2)), list(list(3,4)), list(list(5,6))))
xs <- torch$as_tensor(x$shape)
xp <- x$permute(c(1L, 2L, 0L))
xps <- torch$as_tensor(xp$shape)
print(x) # original tensor
#> tensor([[[1., 2.]],
#>
#> [[3., 4.]],
#>
#> [[5., 6.]]])
print(xp) # permuted tensor
#> tensor([[[1., 3., 5.],
#> [2., 4., 6.]]])
print(xs) # shape original tensor
#> tensor([3, 1, 2])
print(xps) # shape permuted tensor
#> tensor([1, 2, 3])
permute a 3D tensor
torch$manual_seed(1234)
#> <torch._C.Generator>
x <- torch$randn(10L, 480L, 640L, 3L)
x[1:3, 1:2, 1:3, 1:2]
#> tensor([[[[-0.0883, 0.3420],
#> [ 1.0051, -0.1117],
#> [-0.0982, -0.3511]],
#>
#> [[-0.1465, 0.3960],
#> [-1.6878, 0.5720],
#> [ 0.9426, 2.1187]]],
#>
#>
#> [[[ 0.8107, 0.9289],
#> [ 0.4210, -1.5109],
#> [-1.8483, -0.4636]],
#>
#> [[-1.8324, -1.9304],
#> [-2.7020, 0.3491],
#> [ 0.9180, -1.9872]]],
#>
#>
#> [[[ 1.6555, -0.3531],
#> [ 0.4763, 0.8037],
#> [-0.2171, -0.0839]],
#>
#> [[-0.0886, -1.3389],
#> [ 0.7163, -0.9050],
#> [-0.8144, -1.4922]]]])
xs <- torch$as_tensor(x$size()) # torch$tensor(c(10L, 480L, 640L, 3L))
xp <- x$permute(0L, 3L, 1L, 2L) # specify dimensions order
xps <- torch$as_tensor(xp$size()) # torch$tensor(c(10L, 3L, 480L, 640L))
print(xs) # original tensor size
#> tensor([ 10, 480, 640, 3])
print(xps) # permuted tensor size
#> tensor([ 10, 3, 480, 640])
xp[1:3, 1:2, 1:3, 1:2]
#> tensor([[[[-0.0883, 1.0051],
#> [-0.1465, -1.6878],
#> [-0.6429, 0.5577]],
#>
#> [[ 0.3420, -0.1117],
#> [ 0.3960, 0.5720],
#> [ 0.3014, 0.7813]]],
#>
#>
#> [[[ 0.8107, 0.4210],
#> [-1.8324, -2.7020],
#> [ 1.1724, 0.4434]],
#>
#> [[ 0.9289, -1.5109],
#> [-1.9304, 0.3491],
#> [ 0.9901, -1.3630]]],
#>
#>
#> [[[ 1.6555, 0.4763],
#> [-0.0886, 0.7163],
#> [-0.7774, -0.6281]],
#>
#> [[-0.3531, 0.8037],
#> [-1.3389, -0.9050],
#> [-0.7920, 1.3634]]]])
Logical operations
is it equal
(m0 = torch$zeros(3L, 5L))
#> tensor([[0., 0., 0., 0., 0.],
#> [0., 0., 0., 0., 0.],
#> [0., 0., 0., 0., 0.]])
(m1 = torch$ones(3L, 5L))
#> tensor([[1., 1., 1., 1., 1.],
#> [1., 1., 1., 1., 1.],
#> [1., 1., 1., 1., 1.]])
(m2 = torch$eye(3L, 5L))
#> tensor([[1., 0., 0., 0., 0.],
#> [0., 1., 0., 0., 0.],
#> [0., 0., 1., 0., 0.]])
# is m1 equal to m0
print(m1 == m0)
#> tensor([[False, False, False, False, False],
#> [False, False, False, False, False],
#> [False, False, False, False, False]])
print(as_boolean(m1 == m0))
#> tensor([[False, False, False, False, False],
#> [False, False, False, False, False],
#> [False, False, False, False, False]])
is it not equal
# is it not equal
print(m1 != m1)
#> tensor([[False, False, False, False, False],
#> [False, False, False, False, False],
#> [False, False, False, False, False]])
# are both equal
print(m2 == m2)
#> tensor([[True, True, True, True, True],
#> [True, True, True, True, True],
#> [True, True, True, True, True]])
print(as_boolean(m2 == m2))
#> tensor([[True, True, True, True, True],
#> [True, True, True, True, True],
#> [True, True, True, True, True]])
# some are equal, others don't
m1 != m2
#> tensor([[False, True, True, True, True],
#> [ True, False, True, True, True],
#> [ True, True, False, True, True]])
# some are equal, others don't
m0 != m2
#> tensor([[ True, False, False, False, False],
#> [False, True, False, False, False],
#> [False, False, True, False, False]])
as_boolean(m0 != m2)
#> tensor([[ True, False, False, False, False],
#> [False, True, False, False, False],
#> [False, False, True, False, False]])
AND
# AND
m1 & m1
#> tensor([[1, 1, 1, 1, 1],
#> [1, 1, 1, 1, 1],
#> [1, 1, 1, 1, 1]], dtype=torch.uint8)
as_boolean(m1 & m1)
#> tensor([[True, True, True, True, True],
#> [True, True, True, True, True],
#> [True, True, True, True, True]])
OR
# OR
m0 | m2
#> tensor([[1, 0, 0, 0, 0],
#> [0, 1, 0, 0, 0],
#> [0, 0, 1, 0, 0]], dtype=torch.uint8)
# OR
m1 | m2
#> tensor([[1, 1, 1, 1, 1],
#> [1, 1, 1, 1, 1],
#> [1, 1, 1, 1, 1]], dtype=torch.uint8)
as_boolean(m1 | m2)
#> tensor([[True, True, True, True, True],
#> [True, True, True, True, True],
#> [True, True, True, True, True]])
Extract only one logical result with all
# tensor is less than
A <- torch$ones(60000L, 1L, 28L, 28L)
C <- A * 0.5
# is C < A = TRUE
all(torch$lt(C, A))
#> tensor(1, dtype=torch.uint8)
all(C < A)
#> tensor(1, dtype=torch.uint8)
# is A < C = FALSE
all(A < C)
#> tensor(0, dtype=torch.uint8)
greater than
# tensor is greater than
A <- torch$ones(60000L, 1L, 28L, 28L)
D <- A * 2.0
all(torch$gt(D, A))
#> tensor(1, dtype=torch.uint8)
all(torch$gt(A, D))
#> tensor(0, dtype=torch.uint8)
lower than
# tensor is less than or equal
A1 <- torch$ones(60000L, 1L, 28L, 28L)
all(torch$le(A1, A1))
#> tensor(1, dtype=torch.uint8)
all(A1 <= A1)
#> tensor(1, dtype=torch.uint8)
# tensor is greater than or equal
A0 <- torch$zeros(60000L, 1L, 28L, 28L)
all(torch$ge(A0, A0))
#> tensor(1, dtype=torch.uint8)
all(A0 >= A0)
#> tensor(1, dtype=torch.uint8)
all(A1 >= A0)
#> tensor(1, dtype=torch.uint8)
all(A1 <= A0)
#> tensor(0, dtype=torch.uint8)
As a R logical value
# we implement this little function
all_as_boolean <- function(x) {
# convert tensor of 1s and 0s to a unique boolean
as.logical(torch$all(x)$numpy())
}
all_as_boolean(torch$gt(D, A))
#> [1] TRUE
all_as_boolean(torch$gt(A, D))
#> [1] FALSE
all_as_boolean(A1 <= A1)
#> [1] TRUE
all_as_boolean(A1 >= A0)
#> [1] TRUE
all_as_boolean(A1 <= A0)
#> [1] FALSE
Logical NOT
# vector of booleans
all_true <- torch$BoolTensor(list(TRUE, TRUE, TRUE, TRUE))
all_true
#> tensor([True, True, True, True])
# logical NOT
# negate vector with "!"
not_all_true <- !all_true
not_all_true
#> tensor([False, False, False, False])
# a diagonal matrix
diag <- torch$eye(5L)
diag <- diag$to(dtype=torch$uint8) # convert to unsigned integer
diag
#> tensor([[1, 0, 0, 0, 0],
#> [0, 1, 0, 0, 0],
#> [0, 0, 1, 0, 0],
#> [0, 0, 0, 1, 0],
#> [0, 0, 0, 0, 1]], dtype=torch.uint8)
as_boolean(diag)
#> tensor([[ True, False, False, False, False],
#> [False, True, False, False, False],
#> [False, False, True, False, False],
#> [False, False, False, True, False],
#> [False, False, False, False, True]])
# logical NOT
not_diag <- !diag
not_diag
#> tensor([[0, 1, 1, 1, 1],
#> [1, 0, 1, 1, 1],
#> [1, 1, 0, 1, 1],
#> [1, 1, 1, 0, 1],
#> [1, 1, 1, 1, 0]], dtype=torch.uint8)
# and the negation
!not_diag
#> tensor([[1, 0, 0, 0, 0],
#> [0, 1, 0, 0, 0],
#> [0, 0, 1, 0, 0],
#> [0, 0, 0, 1, 0],
#> [0, 0, 0, 0, 1]], dtype=torch.uint8)
as_boolean(!not_diag)
#> tensor([[ True, False, False, False, False],
#> [False, True, False, False, False],
#> [False, False, True, False, False],
#> [False, False, False, True, False],
#> [False, False, False, False, True]])