Kai Li [Wed, 9 Jul 2014 23:50:31 +0000 (07:50 +0800)]
Replace cudaMemcpy with caffe_gpu_memcpy in SyncedMemory per @longjon
Kai Li [Wed, 9 Jul 2014 17:57:16 +0000 (01:57 +0800)]
Implement @Yangqing's solution to copy memory in the SyncedMemory
Kai Li [Wed, 9 Jul 2014 02:12:07 +0000 (10:12 +0800)]
Switch to GPU mode when pointer is move to or from GPU in SyncedMemory
Kai Li [Wed, 9 Jul 2014 01:47:28 +0000 (09:47 +0800)]
Check the GPU mode to decide which memcpy to use
Kai Li [Mon, 7 Jul 2014 07:26:24 +0000 (15:26 +0800)]
Avoid using cudaMemcpy for memcpy when there is no GPU and CUDA driver
Jeff Donahue [Wed, 9 Jul 2014 22:54:35 +0000 (15:54 -0700)]
Merge pull request #661 from jeffdonahue/fix-error-output
Fix bug introduced by warning logs (errors didn't print)
Jeff Donahue [Wed, 9 Jul 2014 22:47:21 +0000 (15:47 -0700)]
fix bug introduced by warning logs: errors didn't print because they
are logged to the warnings file
Jeff Donahue [Wed, 9 Jul 2014 18:37:14 +0000 (11:37 -0700)]
Makefile: cleanup lint/warn report logic (and make the two more consistent)
Evan Shelhamer [Wed, 9 Jul 2014 06:47:36 +0000 (08:47 +0200)]
Merge pull request #649 from jeffdonahue/warnlog
"make warn" to print compilation warnings; "make lint" runs incrementally
Jeff Donahue [Wed, 9 Jul 2014 00:52:49 +0000 (17:52 -0700)]
Output a lint report for every source file linted; use to lint
incrementally
Jeff Donahue [Tue, 8 Jul 2014 22:58:12 +0000 (15:58 -0700)]
Dump compiler warnings to *.warnings.txt; use "make warn" to print them
Jeff Donahue [Tue, 8 Jul 2014 22:06:34 +0000 (15:06 -0700)]
Merge pull request #647 from jeffdonahue/nocleanlinecount
Don't make clean when running linecount
Jeff Donahue [Tue, 8 Jul 2014 22:02:00 +0000 (15:02 -0700)]
Don't make clean when running linecount
Jeff Donahue [Mon, 7 Jul 2014 23:40:50 +0000 (16:40 -0700)]
Merge pull request #614 from ronghanghu/rectangular_pooling
Rectangular pooling
Ronghang Hu [Mon, 7 Jul 2014 19:12:59 +0000 (12:12 -0700)]
added gradient check for non-square pooling
Ronghang Hu [Mon, 7 Jul 2014 18:32:58 +0000 (11:32 -0700)]
fixed style errors
Evan Shelhamer [Sun, 6 Jul 2014 19:47:53 +0000 (12:47 -0700)]
Merge pull request #611 from shelhamer/makefile-config-cxx
Customize compiler setting in Makefile.config
Ronghang Hu [Sat, 5 Jul 2014 15:21:50 +0000 (08:21 -0700)]
add tests for rectangular pooling regions
Ronghang Hu [Sat, 5 Jul 2014 00:29:56 +0000 (17:29 -0700)]
fixing pooling SetUp() to allow default values for stride and pad
Ronghang Hu [Fri, 4 Jul 2014 04:05:23 +0000 (21:05 -0700)]
Update pooling_layer.cu
Replace pad_, kernel_size_, stride_ with pad_h_, pad_w_, kernel_size_h_, kernel_size_w_, stride_h_, stride_w_ to support pooling on rectangle regions.
Ronghang Hu [Fri, 4 Jul 2014 03:42:58 +0000 (20:42 -0700)]
Update pooling_layer.cpp
Replace pad_, kernel_size_, stride_ with pad_h_, pad_w_, kernel_size_h_, kernel_size_w_, stride_h_, stride_w_ to support pooling on rectangle regions.
Ronghang Hu [Fri, 4 Jul 2014 03:33:50 +0000 (20:33 -0700)]
Update vision_layers.hpp
Replace pad_, kernel_size_, stride_ with pad_h_, pad_w_, kernel_size_h_, kernel_size_w_, stride_h_, stride_w_ to support pooling on rectangle regions.
Ronghang Hu [Fri, 4 Jul 2014 03:30:25 +0000 (20:30 -0700)]
Update caffe.proto
Add pad_h, pad_w, kernel_size_h, kernel_size_w, stride_h, stride_w to support pooling on rectangle regions.
Evan Shelhamer [Fri, 4 Jul 2014 01:15:38 +0000 (18:15 -0700)]
customize compiler setting in Makefile.config
in case of issues with default compilers or exotic platforms.
Evan Shelhamer [Fri, 4 Jul 2014 00:29:14 +0000 (17:29 -0700)]
Merge pull request #549 from jamt9000/fix-resize-crop-pil
Make resizing & cropping with PIL work
Evan Shelhamer [Fri, 4 Jul 2014 00:27:37 +0000 (17:27 -0700)]
Merge pull request #555 from shelhamer/uva-memory
Switch to Unified Virtual Address memory copies
Jeff Donahue [Fri, 4 Jul 2014 00:25:51 +0000 (17:25 -0700)]
Merge pull request #602 from kloudkl/layers-in-order
Layers in order
Evan Shelhamer [Sat, 28 Jun 2014 21:05:35 +0000 (14:05 -0700)]
fix casts (static for void*)
Evan Shelhamer [Sat, 28 Jun 2014 20:48:37 +0000 (13:48 -0700)]
reduce caffe_copy to instantiations, split off caffe_memcpy for void*
Evan Shelhamer [Sat, 28 Jun 2014 08:52:49 +0000 (01:52 -0700)]
replace all memset with caffe_set() / caffe_gpu_set()
...except for `SyncedMem` since it has no type.
Evan Shelhamer [Sat, 28 Jun 2014 02:51:33 +0000 (19:51 -0700)]
replace all memcpy by caffe_copy
Evan Shelhamer [Sat, 28 Jun 2014 04:25:36 +0000 (21:25 -0700)]
do all caffe_copy() as UVA mem copy, and drop caffe_gpu_copy()
Do all memory copies by `cudaMemcpy` in UVA mode so that the same
`caffe_copy()` interface works for all transfers.
`cudaMemcpy()` is used in lieu of BLAS copies because they do not
understand UVA.
Drop the now unnecessary `caffe_gpu_copy()` since location of the
pointers is now irrelevant to the interface.
Evan Shelhamer [Sat, 28 Jun 2014 01:38:39 +0000 (18:38 -0700)]
replace softmax cudaMemcpy with caffe_gpu_copy
Evan Shelhamer [Sat, 28 Jun 2014 01:36:48 +0000 (18:36 -0700)]
switch to unified virtual addressing CUDA memcpy
Host / device copies are distinguished by the virtual address of the
pointers instead of explicit memcpy modes.
Evan Shelhamer [Fri, 27 Jun 2014 22:01:02 +0000 (15:01 -0700)]
report UVA in platform test
Evan Shelhamer [Fri, 4 Jul 2014 00:11:20 +0000 (17:11 -0700)]
Merge pull request #609 from jeffdonahue/multiconv
Generalize CONVOLUTION layer to multiple inputs/outputs
Jeff Donahue [Thu, 3 Jul 2014 22:33:12 +0000 (15:33 -0700)]
ConvolutionLayer can take N bottom blobs and N top blobs
Jeff Donahue [Thu, 3 Jul 2014 21:39:35 +0000 (14:39 -0700)]
add EqualNumBottomTopBlobs() property for layers; use in
ConvolutionLayer
Kai Li [Thu, 3 Jul 2014 12:17:41 +0000 (20:17 +0800)]
Organize the loss layers in alphabetical order
Kai Li [Thu, 3 Jul 2014 12:16:54 +0000 (20:16 +0800)]
Arrange the data layers to be in alphabetical order
Kai Li [Thu, 3 Jul 2014 12:15:53 +0000 (20:15 +0800)]
Separate layers relatively independent of images out of vision_layers
Jeff Donahue [Tue, 1 Jul 2014 17:15:29 +0000 (10:15 -0700)]
fix uninitialized variable warnings in tools
yzhuan [Mon, 30 Jun 2014 08:08:55 +0000 (16:08 +0800)]
Update Makefile.config.example
fix some typing error
Evan Shelhamer [Mon, 30 Jun 2014 00:27:50 +0000 (17:27 -0700)]
Merge pull request #545 from jamt9000/im2col-kernel-test
Test for im2col kernel
Evan Shelhamer [Sat, 28 Jun 2014 20:54:58 +0000 (13:54 -0700)]
lint
Sergio Guadarrama [Sat, 28 Jun 2014 17:37:40 +0000 (10:37 -0700)]
Merge pull request #502 from sguada/fix_dropout_backward
Fix dropout backward in TEST phase
Sergio [Sat, 28 Jun 2014 17:22:13 +0000 (10:22 -0700)]
Remove Cuda.major >= 2 check on Dropout test
Sergio [Thu, 19 Jun 2014 01:49:14 +0000 (18:49 -0700)]
Check that pointers are different before copying in caffe_copy and caffe_gpu_copy
Sergio [Sat, 14 Jun 2014 15:39:35 +0000 (08:39 -0700)]
Added test to Dropout to check gradients during Test phase
Sergio [Sat, 14 Jun 2014 01:54:28 +0000 (18:54 -0700)]
Fix var names in Dropout.cu
Sergio [Sat, 28 Jun 2014 01:54:08 +0000 (18:54 -0700)]
Modify Dropout to allow backward pass in TEST phase
Conflicts:
src/caffe/layers/dropout_layer.cpp
src/caffe/layers/dropout_layer.cu
James Thewlis [Fri, 27 Jun 2014 23:49:53 +0000 (00:49 +0100)]
Fix building tests with parallel make
The changes for .cu tests meant that creating
TEST_BUILD_DIR wasn't happening first
Evan Shelhamer [Fri, 27 Jun 2014 21:15:02 +0000 (14:15 -0700)]
Merge pull request #510 from crizCraig/patch-1
Add comment to Makefile.config.example about DEBUG flag issue in OSX
Evan Shelhamer [Fri, 27 Jun 2014 20:46:49 +0000 (13:46 -0700)]
Merge pull request #531 from flickr/dev-top-k-accuracy
Measure accuracy as top-k (default to top-1)
Rob Hess [Tue, 24 Jun 2014 00:22:43 +0000 (17:22 -0700)]
Comment-fix.
Rob Hess [Fri, 20 Jun 2014 21:39:58 +0000 (14:39 -0700)]
Update name of last added param.
Rob Hess [Fri, 20 Jun 2014 21:37:07 +0000 (14:37 -0700)]
Add unit test for accuracy layer.
cypof [Thu, 19 Jun 2014 22:52:10 +0000 (15:52 -0700)]
Next LayerParameter proto id
Rob Hess [Thu, 19 Jun 2014 22:12:56 +0000 (15:12 -0700)]
Use vectors instead of arrays.
Rob Hess [Mon, 16 Jun 2014 20:35:46 +0000 (13:35 -0700)]
Compute top-k accuracy in AccuracyLayer.
Rob Hess [Sat, 14 Jun 2014 02:07:38 +0000 (19:07 -0700)]
Incorporate top_k param into AccuracyLayer and check it's value.
Rob Hess [Sat, 14 Jun 2014 01:16:04 +0000 (18:16 -0700)]
Add parameter for AccuracyLayer in proto.
Evan Shelhamer [Fri, 27 Jun 2014 15:53:52 +0000 (08:53 -0700)]
Merge pull request #554 from shelhamer/nvcc-arch-50
Add latest CUDA arch to fix invalid device function errors on such devices
Evan Shelhamer [Fri, 27 Jun 2014 15:47:42 +0000 (08:47 -0700)]
add latest CUDA arch to fix invalid device function errors
...on devices of that architecture.
James Thewlis [Fri, 27 Jun 2014 09:49:08 +0000 (10:49 +0100)]
Make resizing & cropping with PIL work
Previously it was trying to use undefined variables
James Thewlis [Thu, 26 Jun 2014 17:12:30 +0000 (18:12 +0100)]
Test for im2col kernel
With associated Makefile changes for .cu tests
This tests that the grid-stride loop works for im2col,
using the CPU version as a reference.
Evan Shelhamer [Fri, 27 Jun 2014 02:13:28 +0000 (19:13 -0700)]
Merge pull request #511 from kloudkl/extract_multiple_features
Extract multiple features in a single Forward pass
Jeff Donahue [Thu, 26 Jun 2014 20:50:33 +0000 (16:50 -0400)]
Merge pull request #546 from BVLC/weight-sharing
Weight Sharing
Evan Shelhamer [Thu, 26 Jun 2014 20:05:44 +0000 (13:05 -0700)]
rename layer -> param mapping for clarity
Evan Shelhamer [Thu, 26 Jun 2014 19:49:00 +0000 (12:49 -0700)]
change weight blob field name to param
Jeff Donahue [Mon, 9 Jun 2014 02:53:45 +0000 (19:53 -0700)]
weight sharing
Evan Shelhamer [Thu, 26 Jun 2014 19:18:49 +0000 (12:18 -0700)]
Merge pull request #497 from jeffdonahue/fix-backward-interface
Improve Backward method interface
Jeff Donahue [Mon, 16 Jun 2014 23:37:17 +0000 (16:37 -0700)]
force_backward works properly with non-backproppable things
Jeff Donahue [Tue, 10 Jun 2014 19:38:59 +0000 (12:38 -0700)]
change Backward interface: propagate_down is a vector -- use to fix
long-standing issue with how this is handled in loss layers (esp.
EuclideanLossLayer)
Evan Shelhamer [Thu, 26 Jun 2014 17:25:12 +0000 (10:25 -0700)]
Merge pull request #522 from sguada/accuracy_without_loss
Split accuracy and loss
Evan Shelhamer [Thu, 26 Jun 2014 17:21:19 +0000 (10:21 -0700)]
file SoftmaxWithLoss in with loss layers
Evan Shelhamer [Thu, 26 Jun 2014 16:10:54 +0000 (09:10 -0700)]
Merge pull request #488 from longjon/wall-werror
Build with -Wall -Werror
Evan Shelhamer [Thu, 26 Jun 2014 00:58:38 +0000 (17:58 -0700)]
content ourselves to -Wall without -Werror for now
Evan Shelhamer [Thu, 26 Jun 2014 00:35:40 +0000 (17:35 -0700)]
make clang++ happy on OSX by not linking with pthread
Evan Shelhamer [Thu, 26 Jun 2014 00:32:27 +0000 (17:32 -0700)]
fix test data layer post-lmdb
Evan Shelhamer [Wed, 25 Jun 2014 23:58:03 +0000 (16:58 -0700)]
Merge pull request #478 from kloudkl/cpu_only_tests
CPU only tests
Jonathan L Long [Sat, 14 Jun 2014 02:59:33 +0000 (19:59 -0700)]
turn off some warnings for older compilers
Jonathan L Long [Tue, 10 Jun 2014 22:39:44 +0000 (15:39 -0700)]
add WARNINGS to CXXFLAGS
Jonathan L Long [Tue, 10 Jun 2014 22:38:06 +0000 (15:38 -0700)]
upgrade warnings to -Wall -Werror -Wno-sign-compare
Jonathan L Long [Tue, 10 Jun 2014 22:34:01 +0000 (15:34 -0700)]
don't end comments with \, so that -Wcomment can be used
Jonathan L Long [Tue, 10 Jun 2014 21:50:24 +0000 (14:50 -0700)]
initialize and comment variables that the compiler finds suspicious
Jonathan L Long [Tue, 10 Jun 2014 21:45:36 +0000 (14:45 -0700)]
move CUDA 6.0 check into switch statement itself
This allows -Wswitch to be turned on so that the compiler can check
exhaustiveness.
Jonathan L Long [Wed, 11 Jun 2014 22:29:04 +0000 (15:29 -0700)]
add missing const qualifiers to MemoryDataLayer ExactNum* functions
Jonathan L Long [Thu, 12 Jun 2014 02:31:40 +0000 (19:31 -0700)]
remove unused variables from tests
Jonathan L Long [Mon, 26 May 2014 09:31:34 +0000 (02:31 -0700)]
remove unused variables
Jonathan L Long [Thu, 12 Jun 2014 02:31:22 +0000 (19:31 -0700)]
initialize in declared order in tests
Jonathan L Long [Mon, 26 May 2014 09:31:05 +0000 (02:31 -0700)]
initialize in declared order
Jonathan L Long [Tue, 10 Jun 2014 22:34:52 +0000 (15:34 -0700)]
check if window file is empty in WindowDataLayer
Also note that window files containing windows with different numbers of
channels may not work correctly.
Jonathan L Long [Mon, 26 May 2014 09:02:42 +0000 (02:02 -0700)]
actually check status values from all HDF5 calls
Evan Shelhamer [Wed, 25 Jun 2014 21:47:06 +0000 (14:47 -0700)]
Merge pull request #427 from jamt9000/fix-kernel-index
Should not modify index in im2col kernel loop
Evan Shelhamer [Wed, 25 Jun 2014 03:01:10 +0000 (11:01 +0800)]
fix SOFTMAX_LOSS to work with loss top blob interface
Kai Li [Tue, 24 Jun 2014 14:14:54 +0000 (22:14 +0800)]
Init google logging
Kai Li [Fri, 20 Jun 2014 02:14:19 +0000 (10:14 +0800)]
Replace the raw pointers with shared_ptr to ensure memory is released
Kai Li [Tue, 17 Jun 2014 08:56:30 +0000 (16:56 +0800)]
No need to manually delete the pointers which are managed by std::vector
Kai Li [Tue, 17 Jun 2014 08:27:24 +0000 (16:27 +0800)]
Progress should be reported for each feature blob