Ronghang Hu [Sat, 26 Sep 2015 18:14:30 +0000 (11:14 -0700)]
Merge pull request #3073 from ronghanghu/consistent-malloc-free
Ensure consistency between memory alloc and free
Ronghang Hu [Wed, 16 Sep 2015 20:39:23 +0000 (13:39 -0700)]
Add flag on how host memory is allocated
Add a bool flag to record whether a host memory is allocated using malloc or
cudaMallocHost, and free correspondingly using this flag, instead of depending on Caffe::mode(), which is mutable during runtime.
Evan Shelhamer [Fri, 25 Sep 2015 21:57:17 +0000 (14:57 -0700)]
Merge pull request #3032 from ronghanghu/ban-pythonlayer-in-parallel
Disallow PythonLayer in Multi-GPU training
Evan Shelhamer [Fri, 25 Sep 2015 16:23:32 +0000 (09:23 -0700)]
Merge pull request #3118 from ducha-aiki/patch-1
Fix parse_log.sh against "prefetch queue empty" messages
Dmytro Mishkin [Fri, 25 Sep 2015 07:00:23 +0000 (10:00 +0300)]
Fix parse_log.sh against "prefetch queue empty" messages
Jeff Donahue [Thu, 24 Sep 2015 21:10:50 +0000 (14:10 -0700)]
Merge pull request #3115 from lukeyeager/fix-generate-sample-data
Fix generate_sample_data.py
Luke Yeager [Thu, 24 Sep 2015 20:33:11 +0000 (13:33 -0700)]
Fix generate_sample_data.py - bug from #2978
Jeff Donahue [Thu, 24 Sep 2015 20:11:05 +0000 (13:11 -0700)]
Merge pull request #2978 from lukeyeager/h5t_integer
Allow H5T_INTEGER in HDF5 files
Luke Yeager [Wed, 26 Aug 2015 01:58:45 +0000 (18:58 -0700)]
Modify HDF5DataLayerTest to test H5T_INTEGER data
Luke Yeager [Thu, 24 Sep 2015 19:35:35 +0000 (12:35 -0700)]
Allow H5T_INTEGER in HDF5 files
Evan Shelhamer [Thu, 24 Sep 2015 04:17:45 +0000 (21:17 -0700)]
Merge pull request #2951 from dplarson/examples_readme_link
[examples] fix link to feature visualization notebook
Evan Shelhamer [Thu, 24 Sep 2015 03:38:44 +0000 (20:38 -0700)]
Merge pull request #3112 from shelhamer/test-reshape-harder
[test] Test Reshape more rigorously
Evan Shelhamer [Wed, 23 Sep 2015 20:40:24 +0000 (13:40 -0700)]
[test] TestReshape: check that shapes actually change
Check that output spatial shape varies with input shape while the output
num matches the input num.
Evan Shelhamer [Wed, 23 Sep 2015 20:40:16 +0000 (13:40 -0700)]
[test] TestReshape: expect instead of check
Evan Shelhamer [Wed, 23 Sep 2015 20:38:29 +0000 (13:38 -0700)]
[test] TestReshape: check small then large
checking large then small can mask failure since the smaller shape
memory will fit within the larger shape.
Jon Long [Mon, 21 Sep 2015 23:27:51 +0000 (16:27 -0700)]
Merge pull request #3096 from longjon/fix-conv-reshape
Fix broken conv/deconv reshaping
Evan Shelhamer [Mon, 21 Sep 2015 18:06:03 +0000 (11:06 -0700)]
Merge pull request #3093 from shelhamer/check-xcode-clt-version
[build] check xcode command line tools version >= 6
Jonathan L Long [Sun, 20 Sep 2015 22:31:59 +0000 (15:31 -0700)]
fix broken conv/deconv reshaping caused by reading bottom shape in
LayerSetUp
This also eliminates the extra copying of bottom's shape.
Jeff Donahue [Sun, 20 Sep 2015 22:16:59 +0000 (15:16 -0700)]
Merge pull request #3095 from longjon/fix-deconv-backward
Fix broken DeconvolutionLayer GPU backward caused by ND conv typo
Jonathan L Long [Sun, 20 Sep 2015 21:20:28 +0000 (14:20 -0700)]
fix broken DeconvolutionLayer GPU backward caused by typo
Evan Shelhamer [Sun, 20 Sep 2015 03:38:18 +0000 (20:38 -0700)]
[build] check xcode command line tools version >= 6
future-proof version check for BLAS libraries on OS X
fix #3092
Jon Long [Sat, 19 Sep 2015 02:56:20 +0000 (19:56 -0700)]
Merge pull request #3083 from yanchen036/master
refine format of switch case statement in solver
Jeff Donahue [Sat, 19 Sep 2015 01:13:05 +0000 (18:13 -0700)]
Merge pull request #2049 from jeffdonahue/nd-convolution
ND convolution with im2col
Jeff Donahue [Thu, 5 Mar 2015 03:30:17 +0000 (19:30 -0800)]
Im2col and Convolution layers support N spatial axes
Jeff Donahue [Thu, 5 Mar 2015 05:31:34 +0000 (21:31 -0800)]
Blob: add SyncedMemory shape accessor for GPU shape access
Jeff Donahue [Thu, 5 Mar 2015 03:27:56 +0000 (19:27 -0800)]
caffe.proto: generalize ConvolutionParameter to N spatial axes
Yan Chen [Fri, 18 Sep 2015 09:02:16 +0000 (17:02 +0800)]
refine format of switch case in solver
Evan Shelhamer [Thu, 17 Sep 2015 23:38:39 +0000 (16:38 -0700)]
Merge pull request #3081 from shelhamer/include-io-by-default
include IO dependencies by default
Evan Shelhamer [Thu, 17 Sep 2015 22:05:12 +0000 (15:05 -0700)]
[build] include IO dependencies by default
keep old behavior by including leveldb, lmdb, and opencv by default
Ronghang Hu [Thu, 17 Sep 2015 20:26:30 +0000 (13:26 -0700)]
Merge pull request #3074 from ronghanghu/show-use-cpu
Get back 'USE CPU' print for caffe train
Jeff Donahue [Thu, 17 Sep 2015 07:33:11 +0000 (00:33 -0700)]
Merge pull request #2523 from BonsaiAI/separate-io-dependencies
Separate dependencies for configurable installation
T.E.A de Souza [Tue, 8 Sep 2015 04:20:40 +0000 (12:20 +0800)]
Add a comment indicating that Travis CI tests are CPU only
Tea [Thu, 17 Sep 2015 07:02:45 +0000 (15:02 +0800)]
Fix case in CMake notices
Tea [Sat, 27 Jun 2015 03:44:56 +0000 (11:44 +0800)]
Separate IO dependencies
OpenCV, LMDB, LevelDB and Snappy are made optional via switches
(USE_OPENCV, USE_LMDB, USE_LEVELDB) available for Make and CMake
builds. Since Snappy is a LevelDB dependency, its use is determined by
USE_LEVELDB. HDF5 is left bundled because it is used for serializing
weights and solverstates.
Ronghang Hu [Wed, 16 Sep 2015 19:06:16 +0000 (12:06 -0700)]
Get back 'USE CPU' print for caffe train
Jeff Donahue [Wed, 16 Sep 2015 18:24:27 +0000 (11:24 -0700)]
Merge pull request #3070 from mohomran/resize_nd-image_bug
bug in caffe.io.resize_image when applied to images with 2 or >3 channels
Mohamed Omran [Tue, 15 Sep 2015 15:18:32 +0000 (17:18 +0200)]
removed bug in caffe.io.resize_image when applied to Nd images
Ronghang Hu [Mon, 14 Sep 2015 21:57:18 +0000 (14:57 -0700)]
Merge pull request #3049 from seanbell/check-snapshot-prefix
Check that the snapshot directory is writeable before starting training
Sean Bell [Wed, 9 Sep 2015 16:49:27 +0000 (12:49 -0400)]
Check that the snapshot directory is writeable before starting training
Ronghang Hu [Mon, 14 Sep 2015 17:14:56 +0000 (10:14 -0700)]
Merge pull request #3063 from ronghanghu/use-expectnear-in-eltwiselayertest
Use EXPECT_NEAR in EltwiseLayer test
Ronghang Hu [Mon, 14 Sep 2015 16:54:35 +0000 (09:54 -0700)]
Merge pull request #2941 from thatguymike/MultiGPUDocs
Add some documentation on Multi-GPU support
Ronghang Hu [Mon, 14 Sep 2015 03:46:24 +0000 (20:46 -0700)]
Use EXPECT_NEAR in EltwiseLayer test
Otherwise there seem to be some numerical issues causing BLAS results not exactly same as evaluated results in test code.
Ronghang Hu [Wed, 9 Sep 2015 17:25:08 +0000 (10:25 -0700)]
Merge pull request #3048 from seanbell/minor-string-fix
Minor: missing space in string formatting
Sean Bell [Wed, 9 Sep 2015 16:52:52 +0000 (12:52 -0400)]
Minor: missing space in string formatting
Ronghang Hu [Fri, 4 Sep 2015 21:16:31 +0000 (14:16 -0700)]
disallow PythonLayer in Multi-GPU training
Evan Shelhamer [Fri, 4 Sep 2015 17:36:31 +0000 (14:36 -0300)]
Merge pull request #3024 from danielgordon10/python-solver-fix
[pycaffe] expose all solvers for direct instantiation (although note get_solver)
Daniel Gordon [Fri, 4 Sep 2015 17:12:00 +0000 (10:12 -0700)]
enabling the alternate solvers to be accessed by the python interface
Jeff Donahue [Fri, 4 Sep 2015 05:50:13 +0000 (22:50 -0700)]
Merge pull request #3026 from CDLuminate/patch-1
Fix a typo
Evan Shelhamer [Fri, 4 Sep 2015 05:47:51 +0000 (02:47 -0300)]
Merge pull request #3027 from ronghanghu/fix-accuracy-test
Fix AccuracyLayerTest for per-class accuracy.
Ronghang Hu [Fri, 4 Sep 2015 04:44:45 +0000 (21:44 -0700)]
Fix AccuracyLayerTest for per-class accuracy.
Fix AccuracyLayerTest for per-class accuracy. Previously in #2935, it crashes since the test accuracy is nan (0/0) when a class never appear.
Lumin Zhou [Fri, 4 Sep 2015 04:38:43 +0000 (04:38 +0000)]
Update extract_features.cpp
Evan Shelhamer [Fri, 4 Sep 2015 00:14:40 +0000 (21:14 -0300)]
Merge pull request #3023 from jeffdonahue/param-debug-info-fix
fix Net debug_info to report learnable params
Jeff Donahue [Thu, 3 Sep 2015 23:51:07 +0000 (16:51 -0700)]
Merge pull request #2959 from jeffdonahue/netspec-allow-non-iterable-repeated
NetSpec: don't require lists to specify single-element repeated fields
Jeff Donahue [Sat, 22 Aug 2015 00:29:06 +0000 (17:29 -0700)]
NetSpec: don't require lists to specify single-element repeated fields
Jeff Donahue [Thu, 3 Sep 2015 21:57:19 +0000 (14:57 -0700)]
net.cpp fix debug_info params -> learnable_params
Ronghang Hu [Thu, 3 Sep 2015 20:08:30 +0000 (13:08 -0700)]
Merge pull request #3014 from jeffdonahue/trivial-slice-concat
Allow trivial slice/concat
Jeff Donahue [Fri, 23 Jan 2015 20:52:44 +0000 (12:52 -0800)]
ConcatLayer: allow trivial operation with single bottom Blob
Jeff Donahue [Tue, 7 Oct 2014 18:55:54 +0000 (11:55 -0700)]
SliceLayer: allow trivial operation with single top Blob
Jeff Donahue [Wed, 2 Sep 2015 02:37:41 +0000 (19:37 -0700)]
Merge pull request #2990 from mattdawkins/add-openblas-path
Add extra OpenBLAS include search path
Jeff Donahue [Wed, 2 Sep 2015 00:55:03 +0000 (17:55 -0700)]
Merge pull request #3010 from darrengarvey/fixup-memory-leak-siamese
Fix memory leak in convert_mnist_siamese_data.
Jeff Donahue [Wed, 2 Sep 2015 00:43:28 +0000 (17:43 -0700)]
Merge pull request #3012 from lukeyeager/convert-imageset-logging
Show output from convert_imageset tool
Jeff Donahue [Wed, 2 Sep 2015 00:42:26 +0000 (17:42 -0700)]
Merge pull request #3011 from darrengarvey/fixup-cifar10-example
cifar10: Fix examples by setting snapshot_format.
Jeff Donahue [Wed, 2 Sep 2015 00:41:05 +0000 (17:41 -0700)]
Merge pull request #3009 from darrengarvey/cleanup-documentation-warnings
Fix up documentation errors.
Luke Yeager [Wed, 2 Sep 2015 00:20:37 +0000 (17:20 -0700)]
Show output from convert_imageset tool
Darren Garvey [Tue, 1 Sep 2015 23:54:06 +0000 (00:54 +0100)]
Fix up documentation errors.
Fix some doxygen warnings about an undocumented argument in Blob and
incorrect documentation for SoftmaxWithLossLayer::Forward_cpu().
Darren Garvey [Tue, 1 Sep 2015 19:09:25 +0000 (20:09 +0100)]
cifar10: Fix examples by setting snapshot_format.
Commit
4227828a set the default binary format from HDF5 to BINARYPROTO to
fix #2885. This broke the cifar10 examples which relied on this default.
This commit specifies the snapshot_format explicitly since the rest of the
example relies on this being HDF5.
Jeff Donahue [Tue, 1 Sep 2015 23:25:35 +0000 (16:25 -0700)]
Merge pull request #3008 from darrengarvey/cleanup-cmake-variable
Cleanup: Fixup capitalisation of Caffe_POSTFIX.
Darren Garvey [Mon, 10 Aug 2015 01:16:20 +0000 (02:16 +0100)]
Fix memory leak in convert_mnist_siamese_data.
This fixes a memory leak by using delete[] rather than plain delete.
Darren Garvey [Tue, 1 Sep 2015 14:42:26 +0000 (15:42 +0100)]
Cleanup: Fixup capitalisation of Caffe_POSTFIX.
Replaces CAffe_POSTFIX -> Caffe_POSTFIX.
Jeff Donahue [Tue, 1 Sep 2015 20:42:15 +0000 (13:42 -0700)]
Merge pull request #3007 from philkr/neg_lr_mult
Compute backward for negative lr_mult
philkr [Tue, 1 Sep 2015 20:11:26 +0000 (13:11 -0700)]
Compute backward for negative lr_mult
Jeff Donahue [Sun, 30 Aug 2015 20:43:20 +0000 (13:43 -0700)]
Merge pull request #2998 from longjon/data-race
Fix a recently introduced race condition in DataLayer
Ronghang Hu [Sun, 30 Aug 2015 05:40:12 +0000 (22:40 -0700)]
Merge pull request #2894 from mfigurnov/fix-truncation-warning
Fix truncation of value warning
Jonathan L Long [Sat, 29 Aug 2015 04:27:11 +0000 (21:27 -0700)]
fix GPU data race
Previously, the prefetch GPU -> top GPU and prefetch CPU -> prefetch GPU
copies were launched concurrently in separate streams, allowing the next
batch to be copied in before the current one is read.
This patch explicitly synchronizes the prefetch -> top copy wrt the
host, preventing the CPU -> GPU from being launched until its
completion.
Jeff Donahue [Thu, 27 Aug 2015 16:30:00 +0000 (09:30 -0700)]
Merge pull request #2989 from jyegerlehner/embed-layer-compile-warning
Fix EmbedLayer compiler warning for unused variable.
J Yegerlehner [Thu, 27 Aug 2015 15:47:14 +0000 (10:47 -0500)]
Fix EmbedLayer compiler warning for unused variable.
Matt Dawkins [Thu, 27 Aug 2015 14:51:36 +0000 (10:51 -0400)]
Add extra openblas search path
Jon Long [Thu, 27 Aug 2015 01:12:17 +0000 (18:12 -0700)]
Merge pull request #2944 from philkr/python_layer_param
Give the python layer parameter/weight blobs.
Ronghang Hu [Wed, 26 Aug 2015 19:20:32 +0000 (12:20 -0700)]
Merge pull request #2970 from ronghanghu/spp-fix
Fix SPPLayer top blob num and address `pyramid_height_ == 1`
Ronghang Hu [Mon, 24 Aug 2015 21:46:23 +0000 (14:46 -0700)]
Fix previous mistake on unimplemented top and address pyramid_height_==1 in SPPLayer
also, do nothing in SPPLayer Reshape if already reshaped once and bottom size unchanged
Jeff Donahue [Wed, 26 Aug 2015 18:42:38 +0000 (11:42 -0700)]
Merge pull request #2964 from jyegerlehner/mvn-layer-fixes
Fix MVNLayer
Jeff Donahue [Wed, 26 Aug 2015 18:37:49 +0000 (11:37 -0700)]
Merge pull request #2981 from maaskola/draw-deconvolution
Draw Deconvolution layers like Convolution layers
Jonas Maaskola [Sun, 2 Aug 2015 21:47:12 +0000 (23:47 +0200)]
Draw Deconvolution layers like Convolution layers
J Yegerlehner [Sun, 23 Aug 2015 23:57:16 +0000 (18:57 -0500)]
MVNLayer fixes.
Fix the MVNLayer tests so they actually test what they claim.
MVNLayer fixes: sum_multiplier_ sized correctly; backward gradient calculation.
Gradient calculation per analysis of seanbell, found here:
https://github.com/BVLC/caffe/issues/1938
Fixes according to review comments.
Jeff Donahue [Wed, 26 Aug 2015 01:23:18 +0000 (18:23 -0700)]
Merge pull request #2083 from jeffdonahue/tile-layer
TileLayer
Jeff Donahue [Wed, 19 Aug 2015 01:15:20 +0000 (18:15 -0700)]
TileLayer: add CUDA kernels
Jeff Donahue [Wed, 31 Dec 2014 22:07:00 +0000 (14:07 -0800)]
Add TileLayer
Evan Shelhamer [Tue, 25 Aug 2015 23:24:57 +0000 (20:24 -0300)]
Merge pull request #2032 from jeffdonahue/embed-layer
Embed layer for lookup table of one hot encodings
Evan Shelhamer [Tue, 25 Aug 2015 22:05:57 +0000 (19:05 -0300)]
Merge pull request #2927 from jeffdonahue/improve-net-init-error-msgs
improve net config and shape mismatch error messages
Evan Shelhamer [Tue, 25 Aug 2015 22:03:40 +0000 (19:03 -0300)]
Merge pull request #2972 from jeffdonahue/concat-backward-fix
[fix] properly backprop through ConcatLayer with propagate_down set
philkr [Tue, 25 Aug 2015 17:20:53 +0000 (10:20 -0700)]
Python parameter test added
philkr [Wed, 19 Aug 2015 22:11:30 +0000 (15:11 -0700)]
Allow the python layer have weight/parameter blobs.
Jeff Donahue [Tue, 25 Aug 2015 03:46:38 +0000 (20:46 -0700)]
Merge pull request #2963 from longjon/superfluous-toproto
Remove superfluous code in Net::ToProto
Jeff Donahue [Tue, 25 Aug 2015 02:22:54 +0000 (19:22 -0700)]
bugfix for ConcatLayer with propagate_down set
if propagate_down[i] was set, offset_concat_axis was not correctly
updated for subsequent bottoms i+1, i+2, ...
Jeff Donahue [Tue, 25 Aug 2015 02:44:18 +0000 (19:44 -0700)]
TestConcatLayer: add gradient check for bottom[1] only (to verify
propagate_down[0] == false correctness)
Ronghang Hu [Sat, 22 Aug 2015 23:17:12 +0000 (16:17 -0700)]
Merge pull request #2935 from rmanor/accuracies
Output accuracies per class.
Ronghang Hu [Sat, 22 Aug 2015 23:08:39 +0000 (16:08 -0700)]
Merge pull request #2253 from jyegerlehner/snapshot_on_signal
Snapshot on signal
Ran [Sat, 15 Aug 2015 17:09:43 +0000 (20:09 +0300)]
Output accuracies per class.
Fixed case where number of samples in class can be zero.
- Fixed ignore_label case, also added a test.
- Two other fixes.
Fixed lint errors.
Small fix.
J Yegerlehner [Fri, 3 Apr 2015 21:11:23 +0000 (16:11 -0500)]
Add signal handler and early exit/snapshot to Solver.
Add signal handler and early exit/snapshot to Solver.
Add signal handler and early exit/snapshot to Solver.
Also check for exit and snapshot when testing.
Skip running test after early exit.
Fix more lint.
Rebase on master.
Finish rebase on master.
Fixups per review comments.
Redress review comments.
Lint.
Correct error message wording.
Jonathan L Long [Sat, 22 Aug 2015 04:23:22 +0000 (21:23 -0700)]
remove superfluous code in Net::ToProto