platform/upstream/caffe.git
10 years agoFixed MemoryDataLayer to make it work with pycaffe
Sergio Guadarrama [Thu, 11 Sep 2014 22:19:57 +0000 (15:19 -0700)]
Fixed MemoryDataLayer to make it work with pycaffe

10 years agoMove data_mean into data_transformer remove datum_* from BaseData
Sergio Guadarrama [Sat, 20 Sep 2014 00:10:59 +0000 (17:10 -0700)]
Move data_mean into data_transformer remove datum_* from BaseData

Conflicts:

src/caffe/layers/memory_data_layer.cpp

10 years agoMerge pull request #1204 from jeffdonahue/exp-layer
Jeff Donahue [Fri, 3 Oct 2014 01:26:44 +0000 (18:26 -0700)]
Merge pull request #1204 from jeffdonahue/exp-layer

EXP neuron layer

10 years agosave/restore shared weights unit test
Jeff Donahue [Fri, 3 Oct 2014 00:11:38 +0000 (17:11 -0700)]
save/restore shared weights unit test

10 years ago[fix] comment typo
Jeff Donahue [Thu, 2 Oct 2014 23:20:04 +0000 (16:20 -0700)]
[fix] comment typo

10 years agoAdd ExpLayer to calculate y = base ^ (scale * x + shift)
Jeff Donahue [Thu, 2 Oct 2014 06:12:06 +0000 (23:12 -0700)]
Add ExpLayer to calculate y = base ^ (scale * x + shift)

10 years agoAdd caffe_gpu_exp math function
Jeff Donahue [Thu, 2 Oct 2014 06:11:41 +0000 (23:11 -0700)]
Add caffe_gpu_exp math function

10 years agoMerge pull request #1208 from Yangqing/dev
Yangqing Jia [Thu, 2 Oct 2014 20:01:17 +0000 (13:01 -0700)]
Merge pull request #1208 from Yangqing/dev

add factory header to caffe hpp

10 years agoMerge pull request #1196 from longjon/python-solver-improvements
Evan Shelhamer [Thu, 2 Oct 2014 19:50:46 +0000 (12:50 -0700)]
Merge pull request #1196 from longjon/python-solver-improvements

Python solver improvements

10 years agoadd factory header to caffe hpp
Yangqing Jia [Thu, 2 Oct 2014 19:11:45 +0000 (12:11 -0700)]
add factory header to caffe hpp

10 years agoBack-merge to include #1198
Jonathan L Long [Thu, 2 Oct 2014 18:53:55 +0000 (11:53 -0700)]
Back-merge to include #1198

10 years agoMerge pull request #1198 from dbbert/patch-1
Jonathan L Long [Thu, 2 Oct 2014 18:52:48 +0000 (11:52 -0700)]
Merge pull request #1198 from dbbert/patch-1

10 years agoUpdate detect.py
Bert [Wed, 1 Oct 2014 13:29:24 +0000 (15:29 +0200)]
Update detect.py

This generator causes an error in detect_windows because it is run twice there.
Better store it as a data structure instead.

10 years ago[fix] lint causing travis failures
Jeff Donahue [Thu, 2 Oct 2014 18:19:05 +0000 (11:19 -0700)]
[fix] lint causing travis failures

10 years agoMerge pull request #951 from gcinbis/patch-1
Evan Shelhamer [Thu, 2 Oct 2014 07:12:31 +0000 (00:12 -0700)]
Merge pull request #951 from gcinbis/patch-1

Lock the mex file to avoid MATLAB crashes

10 years agocaffe.proto: do some minor cleanup (fix comments, alphabetization)
Jeff Donahue [Thu, 2 Oct 2014 03:53:31 +0000 (20:53 -0700)]
caffe.proto: do some minor cleanup (fix comments, alphabetization)

10 years agoBack-merge branch 'master' into dev (for fix in PR #1203)
Jeff Donahue [Wed, 1 Oct 2014 21:47:16 +0000 (14:47 -0700)]
Back-merge branch 'master' into dev (for fix in PR #1203)

10 years agoMerge pull request #1203 from jeffdonahue/fix-output-blob-indexing
Jeff Donahue [Wed, 1 Oct 2014 21:46:18 +0000 (14:46 -0700)]
Merge pull request #1203 from jeffdonahue/fix-output-blob-indexing

[fix] solver indexing of output blobs was incorrect for non-singleton outputs

10 years ago[fix] solver indexing of output blobs was incorrect for non-singleton
Jeff Donahue [Wed, 1 Oct 2014 21:19:30 +0000 (14:19 -0700)]
[fix] solver indexing of output blobs was incorrect for non-singleton
outputs

10 years agoMerge pull request #1096 from qipeng/smoothed-cost
Jeff Donahue [Wed, 1 Oct 2014 07:50:55 +0000 (00:50 -0700)]
Merge pull request #1096 from qipeng/smoothed-cost

Display averaged loss over the last several iterations

10 years ago[pycaffe] expose SGDSolver.iter
Jonathan L Long [Wed, 1 Oct 2014 03:26:25 +0000 (20:26 -0700)]
[pycaffe] expose SGDSolver.iter

10 years agoadd accessor for Solver::iter_
Jonathan L Long [Wed, 1 Oct 2014 03:24:47 +0000 (20:24 -0700)]
add accessor for Solver::iter_

This will be useful in code controlling the solver (from pycaffe, for
example).

10 years ago[pycaffe] expose SGDSolver.test_nets
Jonathan L Long [Tue, 30 Sep 2014 23:21:18 +0000 (16:21 -0700)]
[pycaffe] expose SGDSolver.test_nets

10 years ago[pycaffe] add converter for vectors of Nets
Jonathan L Long [Tue, 30 Sep 2014 23:20:27 +0000 (16:20 -0700)]
[pycaffe] add converter for vectors of Nets

10 years ago[pycaffe] fix comment typo
Jonathan L Long [Wed, 1 Oct 2014 02:16:56 +0000 (19:16 -0700)]
[pycaffe] fix comment typo

10 years ago[fix] check and load weight for backward w.r.t. data
Evan Shelhamer [Wed, 1 Oct 2014 00:24:26 +0000 (17:24 -0700)]
[fix] check and load weight for backward w.r.t. data

fix NULL pointer when back-propagating to the data for fixed i.e. not
back-propagated weights. same issue as #1048

10 years agoremove a wrong space in common.hpp
Yangqing Jia [Tue, 30 Sep 2014 22:57:38 +0000 (15:57 -0700)]
remove a wrong space in common.hpp

10 years agofix cuDNN build by readding line deleted in #1167
Jonathan L Long [Tue, 30 Sep 2014 22:28:36 +0000 (15:28 -0700)]
fix cuDNN build by readding line deleted in #1167

10 years agoMerge pull request #1191 from jeffdonahue/hdf5data-cleanup
Jeff Donahue [Tue, 30 Sep 2014 19:52:47 +0000 (12:52 -0700)]
Merge pull request #1191 from jeffdonahue/hdf5data-cleanup

HDF5DataLayer cleanup

10 years agohdf5_load_nd_dataset_helper: check that dataset exists first
Jeff Donahue [Tue, 30 Sep 2014 19:32:34 +0000 (12:32 -0700)]
hdf5_load_nd_dataset_helper: check that dataset exists first

10 years agoHDF5DataLayer: die on failure to open source file
Jeff Donahue [Tue, 30 Sep 2014 19:27:09 +0000 (12:27 -0700)]
HDF5DataLayer: die on failure to open source file

10 years agoDie on failure to load HDF5 data file.
Jeff Donahue [Tue, 30 Sep 2014 18:15:21 +0000 (11:15 -0700)]
Die on failure to load HDF5 data file.

10 years agoCleanup HDF5DataLayer log messages.
Jeff Donahue [Tue, 30 Sep 2014 18:13:59 +0000 (11:13 -0700)]
Cleanup HDF5DataLayer log messages.

-Only use "DLOG" during Forward/Backward for less noise during
training/testing
-Cleanup spacing, capitalization, etc.

10 years agohdf5_data_layer.cpp: fix indentation
Jeff Donahue [Tue, 30 Sep 2014 18:37:55 +0000 (11:37 -0700)]
hdf5_data_layer.cpp: fix indentation

10 years agoMerge pull request #1183 from pluskid/hdf5layer
Jeff Donahue [Tue, 30 Sep 2014 18:06:57 +0000 (11:06 -0700)]
Merge pull request #1183 from pluskid/hdf5layer

10 years agoadded test case to cover new HDF5 behavior
Chiyuan Zhang [Fri, 26 Sep 2014 20:20:48 +0000 (16:20 -0400)]
added test case to cover new HDF5 behavior

10 years agomake HDF5 layer support multiple data output
Chiyuan Zhang [Fri, 26 Sep 2014 15:21:59 +0000 (11:21 -0400)]
make HDF5 layer support multiple data output

10 years agotest_gradient_based_solver.cpp: removed unused typedefs
Jeff Donahue [Tue, 30 Sep 2014 18:00:37 +0000 (11:00 -0700)]
test_gradient_based_solver.cpp: removed unused typedefs

10 years agoMerge pull request #1167 from Yangqing/factory
Yangqing Jia [Tue, 30 Sep 2014 17:35:39 +0000 (10:35 -0700)]
Merge pull request #1167 from Yangqing/factory

Layer factory.

10 years agodoxygen
Yangqing Jia [Tue, 30 Sep 2014 17:20:36 +0000 (10:20 -0700)]
doxygen

10 years agoMerge pull request #1187 from longjon/fix-cudnn-pooling-tests
longjon [Tue, 30 Sep 2014 04:40:23 +0000 (21:40 -0700)]
Merge pull request #1187 from longjon/fix-cudnn-pooling-tests

Fix cuDNN pooling tests

10 years agocuDNN pooling layer tests know that nonzero padding is not supported
Jonathan L Long [Tue, 30 Sep 2014 04:21:03 +0000 (21:21 -0700)]
cuDNN pooling layer tests know that nonzero padding is not supported

After #1172, some cuDNN pooling layer tests were failing due to use of
padding. (Before #1172, these tests were actually testing PoolingLayer
instead of CuDNNPoolingLayer via the fallback.) This commit disables
many tests via commenting, so that they can be easily readded when cuDNN
gains pooling padding support.

10 years agouse method overrides for CuDNNPoolingLayer top blob checking
Jonathan L Long [Tue, 30 Sep 2014 04:20:37 +0000 (21:20 -0700)]
use method overrides for CuDNNPoolingLayer top blob checking

10 years agoMerge pull request #1186 from ashafaei/error-fixes
longjon [Tue, 30 Sep 2014 03:35:02 +0000 (20:35 -0700)]
Merge pull request #1186 from ashafaei/error-fixes

Fixed some errors in layer_factory and cudnn_pooling

10 years agoFixed some errors in layer_factory and cudnn_pooling
Alireza Shafaei [Tue, 30 Sep 2014 03:24:00 +0000 (20:24 -0700)]
Fixed some errors in layer_factory and cudnn_pooling

10 years agoMerge remote-tracking branch 'bvlc/master' into dev
Jeff Donahue [Mon, 29 Sep 2014 21:16:58 +0000 (14:16 -0700)]
Merge remote-tracking branch 'bvlc/master' into dev

Conflicts:
src/caffe/layers/hdf5_data_layer.cpp

10 years agoMerge pull request #1166 from pluskid/master
Jeff Donahue [Mon, 29 Sep 2014 21:14:50 +0000 (14:14 -0700)]
Merge pull request #1166 from pluskid/master

bugfix: HDF5 layer reshape bug

10 years ago[examples] adding class names and deploy version of Flickr Style net
Sergey Karayev [Mon, 29 Sep 2014 09:55:22 +0000 (02:55 -0700)]
[examples] adding class names and deploy version of Flickr Style net

10 years agostatic initialization order fiasco
Yangqing Jia [Mon, 29 Sep 2014 02:16:01 +0000 (19:16 -0700)]
static initialization order fiasco

10 years agoadd explicit declaration - does that help the flakyness?
Yangqing Jia [Mon, 29 Sep 2014 01:11:35 +0000 (18:11 -0700)]
add explicit declaration - does that help the flakyness?

10 years agoadd long-ignored threshold layer
Yangqing Jia [Sun, 28 Sep 2014 23:04:53 +0000 (16:04 -0700)]
add long-ignored threshold layer

10 years agoconsolidate duplicate code
Yangqing Jia [Sun, 28 Sep 2014 22:46:03 +0000 (15:46 -0700)]
consolidate duplicate code

10 years agocmake.
Yangqing Jia [Sun, 28 Sep 2014 07:11:05 +0000 (00:11 -0700)]
cmake.

10 years agoMute factory registration message
Yangqing Jia [Sat, 27 Sep 2014 20:18:27 +0000 (13:18 -0700)]
Mute factory registration message

10 years agoAdd back static library. Using -Wl,--whole-archive will allow us to preserve all...
Yangqing Jia [Sat, 27 Sep 2014 20:17:33 +0000 (13:17 -0700)]
Add back static library. Using -Wl,--whole-archive will allow us to preserve all symbols.

10 years agoPre-lunch fun: add a dynamic library guard test.
Yangqing Jia [Sat, 27 Sep 2014 19:06:10 +0000 (12:06 -0700)]
Pre-lunch fun: add a dynamic library guard test.

10 years agomore docs
Yangqing Jia [Fri, 26 Sep 2014 23:17:04 +0000 (16:17 -0700)]
more docs

10 years agorunning factory.
Yangqing Jia [Fri, 26 Sep 2014 20:58:45 +0000 (13:58 -0700)]
running factory.

10 years agoMerge pull request #1172 from Yangqing/conv_factory
Yangqing Jia [Mon, 29 Sep 2014 00:40:45 +0000 (17:40 -0700)]
Merge pull request #1172 from Yangqing/conv_factory

Pooling factory

10 years agomessage
Yangqing Jia [Sun, 28 Sep 2014 22:33:52 +0000 (15:33 -0700)]
message

10 years agoMerge pull request #1161 from jjkjkj/dev-threshold-fix
Yangqing Jia [Sun, 28 Sep 2014 15:16:52 +0000 (08:16 -0700)]
Merge pull request #1161 from jjkjkj/dev-threshold-fix

Fix Threshold layer

10 years agoconst fix
Yangqing Jia [Sun, 28 Sep 2014 03:22:10 +0000 (20:22 -0700)]
const fix

10 years agocudnn pooling fallback option
Yangqing Jia [Sun, 28 Sep 2014 03:10:13 +0000 (20:10 -0700)]
cudnn pooling fallback option

10 years agofix hdf5 data layer bug
Chiyuan Zhang [Fri, 26 Sep 2014 19:21:57 +0000 (15:21 -0400)]
fix hdf5 data layer bug

10 years agoupdate HDF5 layer test data.
Chiyuan Zhang [Fri, 26 Sep 2014 19:21:45 +0000 (15:21 -0400)]
update HDF5 layer test data.

10 years agotweak test case to expose bug.
Chiyuan Zhang [Fri, 26 Sep 2014 17:18:38 +0000 (13:18 -0400)]
tweak test case to expose bug.

10 years agoUpdate layer_factory.cpp
jjkjkj [Fri, 26 Sep 2014 09:39:49 +0000 (12:39 +0300)]
Update layer_factory.cpp

10 years agoMerge pull request #1149 from ashafaei/crop_bugfix
longjon [Thu, 25 Sep 2014 17:35:32 +0000 (10:35 -0700)]
Merge pull request #1149 from ashafaei/crop_bugfix

Fixed the bug with random offset selection

10 years agoMerge pull request #1157 from ducha-aiki/fix-extract-features
longjon [Thu, 25 Sep 2014 17:22:23 +0000 (10:22 -0700)]
Merge pull request #1157 from ducha-aiki/fix-extract-features

Small fix in extract_features

10 years agoRandom crop bugfix and abstracting random number generation inside data_transformer
Alireza Shafaei [Wed, 24 Sep 2014 05:35:00 +0000 (22:35 -0700)]
Random crop bugfix and abstracting random number generation inside data_transformer

10 years agoRemoved unnecessary "mutable"
D.Mishkin [Thu, 25 Sep 2014 14:30:48 +0000 (17:30 +0300)]
Removed unnecessary "mutable"

10 years agoMerge pull request #1147 from savy-91/patch-1
Evan Shelhamer [Thu, 25 Sep 2014 03:29:23 +0000 (20:29 -0700)]
Merge pull request #1147 from savy-91/patch-1

Changed "blas" to "openblas"

10 years agoMerge pull request #1138 from ksimonyan/vgg_models_support
Evan Shelhamer [Thu, 25 Sep 2014 03:19:39 +0000 (20:19 -0700)]
Merge pull request #1138 from ksimonyan/vgg_models_support

VGG models support

10 years agoBack-merge
Evan Shelhamer [Wed, 24 Sep 2014 21:48:04 +0000 (14:48 -0700)]
Back-merge

  Fixed param order of cv::Size in cv::resize
  switch examples to lmdb (except for custom data loaders)
  fix cifar10 paths so they can be run from caffe root
  default backend to lmdb for image conversion and mean computation

10 years agoMerge pull request #1152 from sguada/fix_cv_size_order
Evan Shelhamer [Wed, 24 Sep 2014 21:43:35 +0000 (14:43 -0700)]
Merge pull request #1152 from sguada/fix_cv_size_order

Fix param order of cv::Size in cv::resize

10 years agoFixed param order of cv::Size in cv::resize
Sergio [Wed, 24 Sep 2014 21:27:57 +0000 (14:27 -0700)]
Fixed param order of cv::Size in cv::resize

10 years agoadded a Matlab demo with mean BGR pixel subtraction instead of the mean image subtraction
Karen Simonyan [Wed, 24 Sep 2014 21:08:46 +0000 (22:08 +0100)]
added a Matlab demo with mean BGR pixel subtraction instead of the mean image subtraction

10 years agoChanged "blas" to "openblas"
savy-91 [Tue, 23 Sep 2014 19:57:16 +0000 (21:57 +0200)]
Changed "blas" to "openblas"

The change was made because openblas has to be linked as "-lopenblas" and not "-lblas". This caused an error when compiling.

10 years agoRGB -> BGR in the matlab demo
Karen Simonyan [Tue, 23 Sep 2014 15:39:08 +0000 (16:39 +0100)]
RGB -> BGR in the matlab demo

10 years agoadded example usage to the Matlab script
Karen Simonyan [Mon, 22 Sep 2014 19:39:37 +0000 (20:39 +0100)]
added example usage to the Matlab script

10 years agoadded comments to the Matlab demo script
Karen Simonyan [Mon, 22 Sep 2014 19:09:35 +0000 (20:09 +0100)]
added comments to the Matlab demo script

10 years agoadded matcaffe_demo for the VGG models (RGB input)
Karen Simonyan [Sun, 21 Sep 2014 16:59:25 +0000 (17:59 +0100)]
added matcaffe_demo for the VGG models (RGB input)

10 years agoadded support for "k" LRN parameter to upgrade_proto
Karen Simonyan [Sun, 21 Sep 2014 16:58:41 +0000 (17:58 +0100)]
added support for "k" LRN parameter to upgrade_proto

10 years agoadds a parameter to the LRN layer (denoted as "k" in [Krizhevsky et al., NIPS 2012])
Karen Simonyan [Thu, 18 Sep 2014 00:08:56 +0000 (01:08 +0100)]
adds a parameter to the LRN layer (denoted as "k" in  [Krizhevsky et al., NIPS 2012])

10 years agoweb demo fix, closes #1002
Sergey Karayev [Mon, 22 Sep 2014 05:53:30 +0000 (22:53 -0700)]
web demo fix, closes #1002

10 years agoMerge pull request #1128 from shelhamer/default-db-lmdb
Evan Shelhamer [Mon, 22 Sep 2014 05:08:08 +0000 (22:08 -0700)]
Merge pull request #1128 from shelhamer/default-db-lmdb

default to lmdb for database storage

10 years agoswitch examples to lmdb (except for custom data loaders)
Evan Shelhamer [Sun, 21 Sep 2014 22:32:03 +0000 (15:32 -0700)]
switch examples to lmdb (except for custom data loaders)

10 years agofix cifar10 paths so they can be run from caffe root
Jeff Donahue [Fri, 19 Sep 2014 19:38:39 +0000 (12:38 -0700)]
fix cifar10 paths so they can be run from caffe root

10 years agodefault backend to lmdb for image conversion and mean computation
Evan Shelhamer [Sun, 21 Sep 2014 22:20:47 +0000 (15:20 -0700)]
default backend to lmdb for image conversion and mean computation

lmdb is 10-15% faster than leveldb although it takes ~1.1x the storage.
This is usually irrelevant in prefetching since both are fast enough,
but more important lmdb allows multiple, concurrent reads for training
and evaluation several models on the same data.

10 years agoBack-merge
Evan Shelhamer [Sun, 21 Sep 2014 21:21:16 +0000 (14:21 -0700)]
Back-merge

  define up-to-date all-in-one model for pascal finetuning
  load transform params in window data layer
  include WindowDataLayer in data param upgrade
  Fix typo in LRN-expression in docs

10 years agoMerge pull request #1126 from shelhamer/window-data-param-upgrade
Sergio Guadarrama [Sun, 21 Sep 2014 20:21:22 +0000 (13:21 -0700)]
Merge pull request #1126 from shelhamer/window-data-param-upgrade

[fix] include WindowDataLayer in data param upgrade

10 years agodefine up-to-date all-in-one model for pascal finetuning
Evan Shelhamer [Sun, 21 Sep 2014 18:38:20 +0000 (11:38 -0700)]
define up-to-date all-in-one model for pascal finetuning

10 years agoload transform params in window data layer
Evan Shelhamer [Sun, 21 Sep 2014 17:29:39 +0000 (10:29 -0700)]
load transform params in window data layer

10 years agoinclude WindowDataLayer in data param upgrade
Evan Shelhamer [Sun, 21 Sep 2014 16:52:51 +0000 (09:52 -0700)]
include WindowDataLayer in data param upgrade

10 years agoMerge pull request #1115 from rickardnorlander/master
Jeff Donahue [Sat, 20 Sep 2014 15:39:14 +0000 (08:39 -0700)]
Merge pull request #1115 from rickardnorlander/master

Fix typo in LRN-expression in docs

10 years agoMerge pull request #1118 from shelhamer/1x1-conv
longjon [Sat, 20 Sep 2014 06:51:25 +0000 (23:51 -0700)]
Merge pull request #1118 from shelhamer/1x1-conv

Optimize 1x1 convolution for Network-in-Network style operation

10 years agocombine col_{data,diff} into single col_buff to halve memory usage
Evan Shelhamer [Sat, 20 Sep 2014 06:12:12 +0000 (23:12 -0700)]
combine col_{data,diff} into single col_buff to halve memory usage

conv forward / backward only need one of the im2col data and diff
at-a-time so consolidating the two saves a lazy allocation.

10 years agoBack-merge
Evan Shelhamer [Sat, 20 Sep 2014 05:31:52 +0000 (22:31 -0700)]
Back-merge

10 years agooptimize 1x1 convolution for Network-in-Network style layers
Evan Shelhamer [Sat, 20 Sep 2014 01:59:38 +0000 (18:59 -0700)]
optimize 1x1 convolution for Network-in-Network style layers

1x1 convolution with stride 1 is a special case of Caffe matrix
multiplication convolution for which im2col / col2im transformations are
actually the identity. For this special case the memory and
transformation are skipped.

10 years agodrop out-of-date conv test comments
Evan Shelhamer [Sat, 20 Sep 2014 04:32:16 +0000 (21:32 -0700)]
drop out-of-date conv test comments