Master/add cudnn support#16

Draft

codethinki wants to merge 20 commits intomasterfrom

_master/add_cudnn_support

Commits on Feb 25, 2026

cleaned up the cmake stuff a little bit, added new features for vcpkg to cover the dependencies automatically (for the most part)
Lukas Thomann
committed
renamed "find" to "dependencies"
Lukas Thomann
committed

Commits on Feb 27, 2026

accidentally deleted cpu and cuda features
Lukas Thomann
committed
began to add first stuff
Lukas Thomann
committed
Merge branch 'master' into _master/add_cudnn_support
codethinki
authored

Commits on Mar 1, 2026

necessary cmake modifications
Lukas Thomann
committed
fully migrated to cudnn v8
Lukas Thomann
committed

Commits on Mar 2, 2026

fixed bug whcih forced type to be float instead of tensor type
Lukas Thomann
committed

Commits on Mar 9, 2026

just having a really hard time fixing the tests and bugs but it'll happen
Lukas Thomann
committed

Commits on Mar 18, 2026

LFG found what triggered cmake reconfigure on test code save
Lukas Thomann
committed

Commits on Mar 19, 2026

still actively working on it just very hard
Lukas Thomann
committed
very small api change (sorry)
Lukas Thomann
committed

Commits on Mar 21, 2026

cv pos fixes & other breadcrumbs
Lukas Thomann
committed
Tensor::concat now supports arbitrarily many tensors
Lukas Thomann
committed
fixed dispatch type not working with R = void
Lukas Thomann
committed
utils "fromFlData" now uses the dispatch type function
Lukas Thomann
committed
minor stuff, still havent fixed the tests
Lukas Thomann
committed

Commits on Mar 24, 2026

small fixes
Lukas Thomann
committed

Commits on Apr 2, 2026

breadcrumbs
Lukas Thomann
committed

Commits on Apr 3, 2026

temp commit for sync
Lukas Thomann
committed