Created release v19.21
dlib/CMakeLists.txt | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
Use a cache to avoid calls to the cuDNN algorithm selection routines.
dlib/cuda/cudnn_dlibapi.cpp | 50 +++++++++++++++++++++++++++------------------ dlib/cuda/cudnn_dlibapi.h | 2 -- 2 files changed, 30 insertions(+), 22 deletions(-)
fix some spelling and grammar errors
dlib/algs.h | 2 +- dlib/bigint/bigint_kernel_abstract.h | 2 +- dlib/disjoint_subsets/disjoint_subsets_sized_abstract.h | 2 +- dlib/gui_core/gui_core_kernel_abstract.h | 2 +- dlib/gui_widgets/base_widgets.cpp | 2 +- dlib/gui_widgets/widgets.h | 2 +- dlib/image_keypoint/binned_vector_feature_image_abstract.h | 2 +- dlib/image_transforms/interpolation_abstract.h | 6 +++--- dlib/lsh/create_random_projection_hash_abstract.h | 4 ++-- dlib/statistics/statistics_abstract.h | 2 +- dlib/svm/structural_svm_distributed_abstract.h | 2 +- dlib/svm/structural_svm_problem_abstract.h | 2 +- tools/htmlify/to_xml.cpp | 2 +- tools/python/src/image2.cpp | 8 ++++---- 14 files changed, 20 insertions(+), 20 deletions(-)
More optimization unit tests
dlib/test/optimization.cpp | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-)
minor cleanup
dlib/image_loader/png_loader.cpp | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
Added possibility to load PNG images from a data buffer. (#2137)
dlib/image_loader/png_loader.cpp | 157 +++++++++++++++++++++++++------- dlib/image_loader/png_loader.h | 32 ++++++- dlib/image_loader/png_loader_abstract.h | 45 +++++++++ 3 files changed, 198 insertions(+), 36 deletions(-)
Switching to what is hopefully a better fix for the following CUDA error
dlib/CMakeLists.txt | 8 -------- dlib/cuda/cuda_dlib.cu | 4 ++-- 2 files changed, 2 insertions(+), 10 deletions(-)
Avoid unnecessairly asking cuDNN which algorithms to use, since this is slow in cuDNN 8.0
dlib/cuda/cudnn_dlibapi.cpp | 31 ++++++++++++++++++++++--------- dlib/cuda/cudnn_dlibapi.h | 2 ++ 2 files changed, 24 insertions(+), 9 deletions(-)
Cleanup cuDNN conv algorithm selection code slightly by moving it into its own function.
dlib/cuda/cudnn_dlibapi.cpp | 248 ++++++++++++++++++++++++-------------------- dlib/cuda/cudnn_dlibapi.h | 2 + 2 files changed, 135 insertions(+), 115 deletions(-)
oops, fixing a weird typo
dlib/global_optimization/find_max_global.h | 2 +- dlib/global_optimization/find_max_global_abstract.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
tweaked cca test thresholds to avoid false positives
dlib/test/cca.cpp | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
Allow forwarding initial function evaluations into find_max_global()
dlib/global_optimization/find_max_global.h | 215 +++++++++++++-------- .../global_optimization/find_max_global_abstract.h | 204 ++++++++++++------- dlib/test/global_optimization.cpp | 5 + 3 files changed, 268 insertions(+), 156 deletions(-)
Apply --expt-relaxed-constexpr to all older versions of cuda.
dlib/CMakeLists.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Fix restoration of MSVC warnings in public headers (#2135)
dlib/geometry/vector.h | 6 ++++-- dlib/matrix/matrix.h | 8 ++++---- 2 files changed, 8 insertions(+), 6 deletions(-)
Switch cuda target architecture from sm_30 to sm_50. I.e. Maxwell instead of Kepler.
dlib/CMakeLists.txt | 2 +- dlib/cmake_utils/test_for_cuda/CMakeLists.txt | 2 +- dlib/cmake_utils/test_for_cudnn/CMakeLists.txt | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-)
Additional documentation for failed dlib::layer<> use. (#2118)
dlib/dnn/core.h | 3 +++ 1 file changed, 3 insertions(+)
Improve the error a user gets when trying to build the pyhton extention on windows but they didn't install visual C++.
tools/python/CMakeLists.txt | 11 +++++++++++ 1 file changed, 11 insertions(+)
Make cuDNN test project failure print a message saying exactly why it failed.
dlib/CMakeLists.txt | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)
Fix random foreset regression not doing quite the right thing.
dlib/random_forest/random_forest_regression.h | 13 +++++++------ dlib/test/random_forest.cpp | 18 +++++++++++++----- 2 files changed, 20 insertions(+), 11 deletions(-)
add support for cudnn 8.0
dlib/cuda/cudnn_dlibapi.cpp | 85 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 85 insertions(+)
fix paths
tools/python/dlib/__init__.py.in | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Make dlib/__init__.py know the paths to cuda that cmake used. Use those paths instead of relying on CUDA_PATH to be set.
tools/python/CMakeLists.txt | 1 + tools/python/dlib/__init__.py | 13 ------------- tools/python/dlib/__init__.py.in | 20 ++++++++++++++++++++ 3 files changed, 21 insertions(+), 13 deletions(-)
work around a bug in nvcc
dlib/CMakeLists.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Oops, put import statements after path adjustment
tools/python/dlib/__init__.py | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
Add an __init__.py file so we can add paths to CUDA libs on windows.
tools/python/CMakeLists.txt | 6 ++---- tools/python/dlib/__init__.py | 11 +++++++++++ tools/python/src/dlib.cpp | 2 +- 3 files changed, 14 insertions(+), 5 deletions(-)
Record last changeset and set PATCH version to 99
dlib/CMakeLists.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Created release v19.20
dlib/CMakeLists.txt | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)