xtensor
xtensor copied to clipboard
xtensor-stack
C++ tensors with broadcasting and lazy computing
Greetings, Recently I developed a package using xtensor as C++ backend and numpy as a python interface, which is usually like @tdegeus done for some of his packages. I also...
I am trying to evaluate the following partial contraction ```none c(e,q,i,j) = a(e,q,i,j,k,l) * b(e,q,l,k) ``` using ```cpp #include #include #include #include int main() { xt::xtensor a = xt::random::rand({1,4,2,2,2,2}); xt::xtensor...
With every successful project, eventually comes this issue 🌟 Certain fields do try to make a reference for software, even if it was not academically published. I personally believe it...
Test case: ```c++ #include #include #include #include int main() { xt::xarray tensor = {1, 2, 3}; tensor = xt::pad(tensor, {1, 1}); std::cout
### Steps to reproduce Build xtensor from `refs/tags/0.24.6` with bound checks on: ``` "-DXTENSOR_ENABLE_ASSERT=ON" "-DXTENSOR_CHECK_DIMENSION=ON" ``` Run tests and observe: ```console cd /build/source/build/test && ./test_xtensor_lib [doctest] doctest version is "2.4.9"...
On Ubuntu Linux with clang 16.0.2 and c++20, I get one failing test. The test `TEST(xinfo, typename)` produces `le` as type name, instead of `double`.
I haven't found anywhere in xtensor such a generic type, i.e., something like ``xtensor_t`` that we could use for ``xt::xtensor``, ``xt::pytensor``, ``xt::rtensor`` or any other container adaptator using the tensor...
version is 0.24.6 ```cpp xt::xarray t{{1, 2}, {3, 4}}; std::vector rep = {2, 3}; auto tb = xt::tile(t, rep); for(auto b : tb.shape()) std::cout
@tdegeus uses xt::eye often in his code. We need to speed it up! The most straightforward way to speed it up is by using the `assign_to(E)` mechanism. We should probably...
After compiling this test with xsimd, valgrind will show `Invalid write of size` errors. See #1962 for further details.
Metadata
Owner
Metadata
C++ tensors with broadcasting and lazy computing