OrgLion-ML
/
dlib
mirror of https://github.com/davisking/dlib.git
1
0
Fork 0
Code Issues Releases Wiki Activity
5,295 Commits 4 Branches 61 Tags 140 MiB
Commit Graph

5295 Commits

Author SHA1 Message Date
Davis King 3124aa0dba Added CPU implementation of sigmoid() and sigmoid_gradient() 2015-11-18 08:52:03 -05:00
Davis King 7c65c8d28a removed cruft 2015-11-18 08:27:24 -05:00
Davis King 7e12a32b45 Added more kinds of affine_transform(), made the solver use affine_transform() so it 
runs on the GPU, and made affine_transform() take only tensors.
 2015-11-18 08:26:48 -05:00
Davis King c8948ee0ca Finished tensor_rand constructor. 2015-11-16 18:29:09 -05:00
Davis King 3d1c22b9f5 Code cleanup 2015-11-16 18:21:37 -05:00
Davis King 51ea50b3bf More batch normalization optimizations. 2015-11-16 18:04:05 -05:00
Davis King 32125dea2f Optimized batch normalization code 2015-11-16 17:28:10 -05:00
Davis King 273a21cf10 More activation layer updates. Now the relu and fc layer objects 
use either CPU or GPU.  Fixed a bug in gemm().
 2015-11-16 08:55:00 -05:00
Davis King db19a781fc Added a check for constexpr 2015-11-13 16:07:32 -05:00
Davis King 5039f0ba5d Changed the API for functions that can operate in-place to a more appropriate 
form.
 2015-11-13 15:52:19 -05:00
Davis King c1433b3dca Upgrade the layer interface so that you can implement layers that operate 
in-place.
 2015-11-13 15:32:14 -05:00
Davis King 694902920d removed print statements 2015-11-13 10:25:00 -05:00
Davis King adc791eadf Added missing requirements 2015-11-13 10:11:44 -05:00
Davis King 74784872fb removed cruft 2015-11-11 18:13:54 -05:00
Davis King 5e1dc62095 updated specs 2015-11-11 18:09:02 -05:00
Davis King e2a2a26a1b Removed unnecessary zero initialization of parameter gradients in core.h. 2015-11-11 17:59:40 -05:00
Davis King eada4be8e3 Clarified specs 2015-11-11 17:13:34 -05:00
Davis King 8cb8690d72 clarified spec 2015-11-10 08:44:16 -05:00
Davis King a79e7cf645 fixed compile time bugs 2015-11-09 20:25:35 -05:00
Davis King cb2f9de634 Added part of the tensor_tools implementations 2015-11-09 18:29:56 -05:00
Davis King 764338586e Renamed conv to tensor_conv 2015-11-09 18:18:51 -05:00
Davis King 9cbbe7b75b Made a tensor iterable 2015-11-09 17:56:25 -05:00
Davis King 00e18a8d59 Added tests for some of the new tensor operators. 2015-11-09 17:37:15 -05:00
Davis King 1d35a89653 Added the header containing the prototypes that call into either the CPU or GPU 
code based on how dlib was built.  All the layer implementations will interact
with these functions.
 2015-11-09 17:29:41 -05:00
Davis King bc2f02d769 Implemented the CPU side of the tools in cuda_dlib.h and also cleaned some 
things up a bit.
 2015-11-09 17:28:47 -05:00
Davis King 2bc5202b73 fixed some compiler warnings 2015-11-08 17:22:49 -05:00
Davis King 5ae7a4734d Added batch normalization tests 2015-11-08 17:18:46 -05:00
Davis King 6c05ff4558 Added CPU version of batch normalization functions 2015-11-08 16:55:26 -05:00
Davis King 141b384b48 Added binding to cuRAND 2015-11-08 09:34:26 -05:00
Davis King 6539ea6779 Added prototypes for batch normalization's gradients. 2015-11-07 19:59:31 -05:00
Davis King 51ebcfc7f4 Said backward() is allowed to reuse computations cached during the forward() computation. 2015-11-06 11:32:58 -05:00
Davis King 91f512eca4 clarified spec 2015-11-06 09:23:37 -05:00
Davis King 87b182b5f3 merged 2015-11-04 18:06:31 -05:00
Davis King 92073c2003 Cleaned up the test_layer() interface. 2015-11-04 09:06:32 -05:00
Davis King 24698f87ce changed test to avoid compiler error in gcc 4.8 2015-11-03 22:05:18 -05:00
Davis King 3406a290bc Minor tweaks to avoid compiler warnings 2015-11-03 21:48:38 -05:00
Davis King 0145ca8a38 Added a cuBLAS test 2015-11-03 21:38:44 -05:00
Davis King 93410af39e Fixed a bug in cuda::gemm() 2015-11-03 21:37:18 -05:00
Davis King dae8929a64 Added cuda::gemm() 2015-11-03 21:24:59 -05:00
Davis King 7fb29dae6d merged 2015-11-03 18:32:05 -05:00
Davis King 347ee396ba fixed tabbing 2015-11-03 18:31:09 -05:00
Davis E. King 27dd8f6b0a Merge pull request #54 from vrichter/master 
Change relative path of configure files in CMakeLists.
 2015-11-03 18:27:23 -05:00
Davis King e2e26f3bec Relaxed the preconditions so that layers are allowed to output tensors 
that contain a different number of samples than their input tensors.
 2015-11-03 18:15:33 -05:00
Viktor Richter b1dab227dc Change relative path of configure files in CMakeLists. 
Using CMAKE_PROJECT_DIR instead of CMAKE_SOURCE_DIR/../dlib
makes the build process more portable.
 2015-11-03 14:44:30 +01:00
Davis King fd867dd8b9 updated docs 2015-11-02 17:35:59 -05:00
Davis King e98bade3ed Added an MNIST loader 2015-11-02 17:29:04 -05:00
Davis King 074d4b6e15 merged 2015-10-31 07:36:47 -04:00
Davis King 8fcf1a5fda Record last changeset and set PATCH version to 99 2015-10-28 20:27:41 -04:00
Davis King 8badddce46 Added tag v18.18 for changeset 7ae1775f61a4 2015-10-28 20:25:59 -04:00
Davis King 6124442f86 Created release v18.18 2015-10-28 20:25:59 -04:00