Specify integer size for tensor_layout

examples/1_SimpleNet/simplenet_infer_fortran.f90

@@ -2,6 +2,7 @@ program inference
 
    ! Import precision info from iso
    use, intrinsic :: iso_fortran_env, only : sp => real32
+   use, intrinsic :: iso_c_binding, only : c_int64_t
 
    ! Import our library for interfacing with PyTorch
    use ftorch, only : torch_model, torch_tensor, torch_kCPU, torch_delete, &
@@ -22,7 +23,7 @@ program inference
    real(wp), dimension(5), target :: in_data
    real(wp), dimension(5), target :: out_data
    real(wp), dimension(5), target :: expected
-   integer, parameter :: tensor_layout(1) = [1]
+   integer(c_int64_t), parameter :: tensor_layout(1) = [1]
 
    ! Set up Torch data structures
    ! The net, a vector of input tensors (in this case we only have one), and the output tensor

examples/2_ResNet18/resnet_infer_fortran.f90

@@ -1,6 +1,7 @@
 program inference
 
    use, intrinsic :: iso_fortran_env, only : sp => real32
+   use, intrinsic :: iso_c_binding, only : c_int64_t
 
    ! Import our library for interfacing with PyTorch
    use ftorch, only : torch_model, torch_tensor, torch_kCPU, torch_delete, &
@@ -32,10 +33,10 @@ subroutine main()
 
       integer, parameter :: in_dims = 4
       integer, parameter :: in_shape(in_dims) = [1, 3, 224, 224]
-      integer, parameter :: in_layout(in_dims) = [1, 2, 3, 4]
+      integer(c_int64_t), parameter :: in_layout(in_dims) = [1, 2, 3, 4]
       integer, parameter :: out_dims = 2
       integer, parameter :: out_shape(out_dims) = [1, 1000]
-      integer, parameter :: out_layout(out_dims) = [1, 2]
+      integer(c_int64_t), parameter :: out_layout(out_dims) = [1, 2]
 
       ! Path to input data
       character(len=100) :: data_dir

examples/3_MultiGPU/simplenet_infer_fortran.f90

@@ -2,6 +2,7 @@ program inference
 
    ! Import precision info from iso
    use, intrinsic :: iso_fortran_env, only : sp => real32
+   use, intrinsic :: iso_c_binding, only : c_int64_t
 
    ! Import our library for interfacing with PyTorch
    use ftorch, only : torch_model, torch_tensor, torch_kCUDA, torch_kCPU, &
@@ -22,7 +23,7 @@ program inference
    ! Set up Fortran data structures
    real(wp), dimension(5), target :: in_data
    real(wp), dimension(5), target :: out_data
-   integer, parameter :: tensor_layout(1) = [1]
+   integer(c_int64_t), parameter :: tensor_layout(1) = [1]
 
    ! Set up Torch data structures
    type(torch_model) :: model

examples/4_MultiIO/multiionet_infer_fortran.f90

@@ -2,6 +2,7 @@ program inference
 
    ! Import precision info from iso
    use, intrinsic :: iso_fortran_env, only : sp => real32
+   use, intrinsic :: iso_c_binding, only : c_int64_t
 
    ! Import our library for interfacing with PyTorch
    use ftorch, only : torch_model, torch_tensor, torch_kCPU, &
@@ -25,7 +26,7 @@ program inference
    real(wp), dimension(4), target :: out_data1
    real(wp), dimension(4), target :: out_data2
    real(wp), dimension(4) :: expected
-   integer, parameter :: tensor_layout(1) = [1]
+   integer(c_int64_t), parameter :: tensor_layout(1) = [1]
 
    ! Set up Torch data structures
    ! The net, a vector of input tensors (in this case we only have one), and the output tensor

examples/6_Autograd/autograd.f90

@@ -2,6 +2,7 @@ program example
 
   ! Import precision info from iso
   use, intrinsic :: iso_fortran_env, only : sp => real32
+  use, intrinsic :: iso_c_binding, only : c_int64_t
 
   ! Import our library for interfacing with PyTorch's Autograd module
   use ftorch, only : torch_tensor, torch_kCPU, &
@@ -20,7 +21,7 @@ program example
   real(wp), dimension(n,m), target :: in_data
   real(wp), dimension(:,:), pointer :: out_data
   real(wp), dimension(n,m) :: expected
-  integer :: tensor_layout(2) = [1, 2]
+  integer(c_int64_t) :: tensor_layout(2) = [1, 2]
   integer :: i, j
 
   ! Flag for testing

src/ftorch.fypp

@@ -574,7 +574,7 @@ contains
 
     ! inputs
     ${f_type(PREC)}$(kind=${PREC}$), intent(in), target :: data_in${ranksuffix(RANK)}$   !! Input data that tensor will point at
-    integer, intent(in)        :: layout(${RANK}$) !! Control order of indices
+    integer(c_int64_t), intent(in) :: layout(${RANK}$) !! Control order of indices
     integer(c_int), intent(in) :: c_device_type    !! Device type the tensor will live on (`torch_kCPU` or `torch_kCUDA`)
     integer(c_int), optional, intent(in) :: device_index    !! device index to use for `torch_kCUDA` case
     logical, optional, intent(in) :: requires_grad  !! Whether gradients need to be computed for the created tensor

src/test/unit/test_constructors.pf

@@ -16,20 +16,20 @@ subroutine test_torch_tensor_empty()
 
   type(torch_tensor) :: tensor
   integer(c_int), parameter :: ndims = 2
-  integer(c_int64_t), dimension(2) :: tensor_shape
+  integer(c_int64_t), dimension(2) :: tensor_layout
   integer(c_int), parameter :: dtype = torch_kFloat32
   integer(c_int), parameter :: device_type = torch_kCPU
   integer(c_int), parameter :: device_index = -1
   logical(c_bool), parameter :: requires_grad = .false.
   logical :: test_pass
 
-  tensor_shape = [2, 3]
+  tensor_layout = [2, 3]
 
   ! Check the tensor pointer is not associated
   @assertFalse(c_associated(tensor%p))
 
   ! Create a tensor of zeros
-  call torch_tensor_empty(tensor, ndims, tensor_shape, dtype, device_type, &
+  call torch_tensor_empty(tensor, ndims, tensor_layout, dtype, device_type, &
                           device_index, requires_grad)
 
   ! Check the tensor pointer is associated
@@ -52,7 +52,7 @@ subroutine test_torch_tensor_zeros()
 
   type(torch_tensor) :: tensor
   integer(c_int), parameter :: ndims = 2
-  integer(c_int64_t), dimension(2) :: tensor_shape
+  integer(c_int64_t), dimension(2) :: tensor_layout
   integer(c_int), parameter :: dtype = torch_kFloat32
   integer(c_int), parameter :: device_type = torch_kCPU
   integer(c_int), parameter :: device_index = -1
@@ -61,13 +61,13 @@ subroutine test_torch_tensor_zeros()
   real(kind=real32), dimension(2,3) :: expected
   logical :: test_pass
 
-  tensor_shape = [2, 3]
+  tensor_layout = [2, 3]
 
   ! Check the tensor pointer is not associated
   @assertFalse(c_associated(tensor%p))
 
   ! Create a tensor of zeros
-  call torch_tensor_zeros(tensor, ndims, tensor_shape, dtype, device_type, &
+  call torch_tensor_zeros(tensor, ndims, tensor_layout, dtype, device_type, &
                           device_index, requires_grad)
 
   ! Check the tensor pointer is associated
@@ -99,7 +99,7 @@ subroutine test_torch_tensor_ones()
 
   type(torch_tensor) :: tensor
   integer(c_int), parameter :: ndims = 2
-  integer(c_int64_t), dimension(2) :: tensor_shape
+  integer(c_int64_t), dimension(2) :: tensor_layout
   integer(c_int), parameter :: dtype = torch_kFloat32
   integer(c_int), parameter :: device_type = torch_kCPU
   integer(c_int), parameter :: device_index = -1
@@ -108,13 +108,13 @@ subroutine test_torch_tensor_ones()
   real(kind=real32), dimension(2,3) :: expected
   logical :: test_pass
 
-  tensor_shape = [2, 3]
+  tensor_layout = [2, 3]
 
   ! Check the tensor pointer is not associated
   @assertFalse(c_associated(tensor%p))
 
   ! Create tensor of ones
-  call torch_tensor_ones(tensor, ndims, tensor_shape, dtype, device_type, &
+  call torch_tensor_ones(tensor, ndims, tensor_layout, dtype, device_type, &
                          device_index, requires_grad)
 
   ! Check the tensor pointer is associated