Commits · e1e4d6df3a48b909065c86abde754a00f7b0529f · OpenDAS / dlib

"tools/vscode:/vscode.git/clone" did not exist on "ea2fbcab5c3853593712bb7ab962dcbff3971d0b"

21 Nov, 2015 7 commits
- Added cuda implementations of a bunch of functions. · e1e4d6df
  Davis King authored Nov 21, 2015
  
  e1e4d6df
- Added more stuff to tests · ca11ff49
  Davis King authored Nov 21, 2015
  
  ca11ff49
- Made loss layers output the gradients by assigning them to the output rather · 5f5c46f4
  Davis King authored Nov 21, 2015
```
than adding them.  This way, the gradient buffer can be used as scratch space
during the loss computation.
```
  5f5c46f4
- Added more gemm() tests · e2a67dec
  Davis King authored Nov 21, 2015
  
  e2a67dec
- Fixed cmake lists so it compiles correctly without C++11 · fb6c59e4
  Davis King authored Nov 21, 2015
  
  fb6c59e4
- Fixed minor bug · 1717605b
  Davis King authored Nov 21, 2015
  
  1717605b
- Changed gemm so it doesn't cause device to host copies in the assert statements. · c79ae7d4
  Davis King authored Nov 21, 2015
  
  c79ae7d4
20 Nov, 2015 5 commits
- fixed compiler errors · d059f3c1
  Davis King authored Nov 20, 2015
  
  d059f3c1
- Added batch normalization layer skeleton · b084076e
  Davis King authored Nov 20, 2015
  
  b084076e
- Removed cruft · 627db9e1
  Davis King authored Nov 20, 2015
  
  627db9e1
- Added sigmoid, tanh, and softmax layers. Also added log loss layer. · 3bb2a817
  Davis King authored Nov 20, 2015
  
  3bb2a817
- Added log1pexp() · e89d468b
  Davis King authored Nov 20, 2015
  
  e89d468b
19 Nov, 2015 1 commit
- Fixed a compile time bug. · 5e92f10f
  Davis King authored Nov 19, 2015
  
  5e92f10f
18 Nov, 2015 5 commits
- Implemented CPU version of tanh · a7ea7d00
  Davis King authored Nov 18, 2015
  
  a7ea7d00
- Implemented the CPU version of softmax · 9b36bb98
  Davis King authored Nov 18, 2015
  
  9b36bb98
- Added CPU implementation of sigmoid() and sigmoid_gradient() · 3124aa0d
  Davis King authored Nov 18, 2015
  
  3124aa0d
- removed cruft · 7c65c8d2
  Davis King authored Nov 18, 2015
  
  7c65c8d2
- Added more kinds of affine_transform(), made the solver use affine_transform() so it · 7e12a32b
  Davis King authored Nov 18, 2015
```
runs on the GPU, and made affine_transform() take only tensors.
```
  7e12a32b
16 Nov, 2015 5 commits
- Finished tensor_rand constructor. · c8948ee0
  Davis King authored Nov 16, 2015
  
  c8948ee0
- Code cleanup · 3d1c22b9
  Davis King authored Nov 16, 2015
  
  3d1c22b9
- More batch normalization optimizations. · 51ea50b3
  Davis King authored Nov 16, 2015
  
  51ea50b3
- Optimized batch normalization code · 32125dea
  Davis King authored Nov 16, 2015
  
  32125dea
- More activation layer updates. Now the relu and fc layer objects · 273a21cf
  Davis King authored Nov 16, 2015
```
use either CPU or GPU.  Fixed a bug in gemm().
```
  273a21cf
13 Nov, 2015 5 commits
- Added a check for constexpr · db19a781
  Davis King authored Nov 13, 2015
  
  db19a781
- Changed the API for functions that can operate in-place to a more appropriate · 5039f0ba
  Davis King authored Nov 13, 2015
```
form.
```
  5039f0ba
- Upgrade the layer interface so that you can implement layers that operate · c1433b3d
  Davis King authored Nov 13, 2015
```
in-place.
```
  c1433b3d
- removed print statements · 69490292
  Davis King authored Nov 13, 2015
  
  69490292
- Added missing requirements · adc791ea
  Davis King authored Nov 13, 2015
  
  adc791ea
11 Nov, 2015 4 commits
- removed cruft · 74784872
  Davis King authored Nov 11, 2015
  
  74784872
- updated specs · 5e1dc620
  Davis King authored Nov 11, 2015
  
  5e1dc620
- Removed unnecessary zero initialization of parameter gradients in core.h. · e2a2a26a
  Davis King authored Nov 11, 2015
  
  e2a2a26a
- Clarified specs · eada4be8
  Davis King authored Nov 11, 2015
  
  eada4be8
10 Nov, 2015 2 commits
- clarified spec · 8cb8690d
  Davis King authored Nov 10, 2015
  
  8cb8690d
- fixed compile time bugs · a79e7cf6
  Davis King authored Nov 09, 2015
  
  a79e7cf6
09 Nov, 2015 6 commits
- Added part of the tensor_tools implementations · cb2f9de6
  Davis King authored Nov 09, 2015
  
  cb2f9de6
- Renamed conv to tensor_conv · 76433858
  Davis King authored Nov 09, 2015
  
  76433858
- Made a tensor iterable · 9cbbe7b7
  Davis King authored Nov 09, 2015
  
  9cbbe7b7
- Added tests for some of the new tensor operators. · 00e18a8d
  Davis King authored Nov 09, 2015
  
  00e18a8d
- Added the header containing the prototypes that call into either the CPU or GPU · 1d35a896
  Davis King authored Nov 09, 2015
```
code based on how dlib was built.  All the layer implementations will interact
with these functions.
```
  1d35a896
- Implemented the CPU side of the tools in cuda_dlib.h and also cleaned some · bc2f02d7
  Davis King authored Nov 09, 2015
```
things up a bit.
```
  bc2f02d7