Dmitry Kurtaev
d891e9b1d8
Layers for MobileNet from TensorFlow
7 years ago
Dmitry Kurtaev
8646d5fb84
FP16 Caffe models import and export
7 years ago
Dmitry Kurtaev
0ce7c33bc8
Torch's Concat and ConcatTable doesn't use Split layer
7 years ago
Dmitry Kurtaev
7dc6b1d7d4
Layers for OpenFace face recognition network
7 years ago
Dmitry Kurtaev
58b890b9f7
Dilated convolution import from TensorFlow
7 years ago
Dmitry Kurtaev
6639348c5e
Disable SSD on CPU Halide test
7 years ago
Dmitry Kurtaev
ce41a15437
Import and convert FP16 weights from TensorFlow
7 years ago
Maksim Shabunin
248e2c7d47
Fixed some issues found by static analysis
7 years ago
Dmitry Kurtaev
cad7c4d51d
MobileNet-SSD and VGG-SSD topologies in Halide
7 years ago
Alexander Alekhin
01519313d7
dnn: invalid bindings
7 years ago
Dmitry Kurtaev
5c43a394c5
Added performance test for Caffe framework
7 years ago
Aleksandr Rybnikov
8b1146deb2
Added function to get timings for layers
7 years ago
Dmitry Kurtaev
e74c475162
Removed 3rdparty serialized NumPy arrays parser
7 years ago
Dmitry Kurtaev
54f0616a13
Deconvolution layer from TensorFlow
7 years ago
Dmitry Kurtaev
4e28c00e7b
Fix Halide buffer behavior in case of OpenCL device memory allocation
7 years ago
dkurt
339793143c
Unit tests for TensorFlow importer
7 years ago
dkurt
b1ef44b1ac
Replace halide_opencl_device_interface
7 years ago
Aleksandr Rybnikov
8d6b8b45b6
Added ELU and test for it
7 years ago
Ismo Puustinen
c2de5cf735
dnn: force floating point literals to be float.
...
In OpenCL code in activations.cl, make the type of floating point
literals to be float. Otherwise the values will be interpreted as
doubles, causing Beignet to have type conversion issues.
7 years ago
Aleksandr Rybnikov
ce1cc352d9
MobileNet SSD sample
7 years ago
Alexander Alekhin
cbced23de4
cmake: don't include protobuf on disabled DNN module
7 years ago
Alexander Alekhin
3f102e5d3a
dnn: protobuf shutdown
7 years ago
LaurentBerger
69fd57bc96
update tutorial link for dnn
7 years ago
Alexander Alekhin
878a6906cc
dnn: fix torch importer memory leaks
7 years ago
dkurt
70ff3804e9
Fix SqueezeNet Halide performance test
7 years ago
Tomoaki Teshima
0f91faddae
fix linker error when trying CPU_BASELINE=AVX
7 years ago
Tomoaki Teshima
e63d628677
remove some rand functions
...
* make test more reproducible
7 years ago
Alexander Alekhin
08c94aa5c0
build: reuse int32_t workaround from softfloat.hpp
7 years ago
Tomoaki Teshima
1989bc33a7
fix build error on Visual Studio 2012
7 years ago
Aleksandr Rybnikov
7d1140340e
Rewrote googlenet tests
7 years ago
Vadim Pisarevsky
0488d9bdb2
optimize out scaleLayer & concatLayer whenever possible
...
fixed problem in concat layer by disabling memory re-use in layers with multiple inputs
trying to fix the tests when Halide is used to run deep nets
another attempt to fix Halide tests
see if the Halide tests will pass with concat layer fusion turned off
trying to fix failures in halide tests; another try
one more experiment to make halide_concat & halide_enet tests pass
continue attempts to fix halide tests
moving on
uncomment parallel concat layer
seemingly fixed failures in Halide tests and re-enabled concat layer fusion; thanks to dkurt for the patch
7 years ago
Alexander Alekhin
4784c7be5f
dnn: cleanup dispatched code, fix SIMD128 types
7 years ago
Alexander Alekhin
c3e6de293f
dnn: code cleanup, refactor detection output layer
7 years ago
Alexander Alekhin
544908d06c
dnn: some minor fixes in docs, indentation, unused code
7 years ago
dkurt
3203635765
Eltwise layer fixes
7 years ago
Vadim Pisarevsky
ed9564106c
reuse AVX2-optimized kernels for AVX1 CPUs (like IvyBridge)
7 years ago
Aleksandr Rybnikov
d3c14ce3ea
Added tests for ResNet-50 and SqueezeNet v1.1
7 years ago
Alexander Alekhin
32d1eb36df
dnn: fix links
7 years ago
berak
34c0ab5878
dnn: fix some tutorial links
7 years ago
abratchik
8f7181429f
add java wrappers to dnn module
7 years ago
Alexander Alekhin
6ea6e4bceb
binding: fix headers processing
8 years ago
Maksim Shabunin
e0393f8557
Fixed some issues found by static analysis (4th round)
8 years ago
Aleksandr Rybnikov
fab4f4b9d5
Disabled logging in caffe parser in release
8 years ago
Alexander Alekhin
72a765d766
dnn: fix compilation of Halide tests
8 years ago
Alexander Alekhin
511e50c19c
dnn: cleanup torch integration code
8 years ago
Alexander Alekhin
da0960321b
dnn: added "hidden" experimental namespace
...
Main purpose of this namespace is to avoid using of incompatible
binaries that will cause applications crashes.
This additional namespace will not impact "Source code API".
This change allows to maintain ABI checks (with easy filtering out).
8 years ago
Maksim Shabunin
a769d69a9d
Fixed several issues found by static analysis
8 years ago
dkurt
b46f5b1b38
Align convolutional layer weights separately from origin ones
8 years ago
Alexander Alekhin
ed10383359
dnn: added trace macros
8 years ago
Aleksandr Rybnikov
ec321e651f
Removed usage of std::map in DetectionOutput layer
8 years ago