From 4324ab75b302acb16c1a998f51f21a3ac916ff62 Mon Sep 17 00:00:00 2001 From: kchristin Date: Wed, 20 Nov 2024 14:37:01 +0200 Subject: [PATCH] Fix formatting of CUDA doc page --- docs/userDocs/source/user/UsingCladOnCUDACode.rst | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/docs/userDocs/source/user/UsingCladOnCUDACode.rst b/docs/userDocs/source/user/UsingCladOnCUDACode.rst index 49372991d..0be853937 100644 --- a/docs/userDocs/source/user/UsingCladOnCUDACode.rst +++ b/docs/userDocs/source/user/UsingCladOnCUDACode.rst @@ -11,6 +11,7 @@ if not explicitly specified. Note that either none of these two arguments or bot Clad does not handle cases where only one of the two is provided, even if the order is correct. .. code-block:: cpp + #include "clad/Differentiator/Differentiator.h" auto kernel_grad = clad::gradient(kernel, "in, out"); // compute the derivative of out w.r.t in @@ -30,7 +31,8 @@ Clad supports the following CUDA features: To use CUDA math functions, the user must define the equivalent pullback function in Clad's CUDA custom derivatives: .. code-block:: cpp - // *In `clad/include/clad/Differentiator/BuiltinDerivativesCUDA.cuh`* + + // In `clad/include/clad/Differentiator/BuiltinDerivativesCUDA.cuh` namespace clad {