platform/upstream/caffeonacl.git
9 years agoMerge pull request #1473 from longjon/pytest
Evan Shelhamer [Thu, 22 Jan 2015 08:29:35 +0000 (00:29 -0800)]
Merge pull request #1473 from longjon/pytest

Python testing

9 years agohdf5_save_nd_dataset takes a const string& (instead of const string)
Jeff Donahue [Sun, 2 Nov 2014 09:56:19 +0000 (01:56 -0800)]
hdf5_save_nd_dataset takes a const string& (instead of const string)

9 years agoBack-merge fixes + docs
Evan Shelhamer [Wed, 21 Jan 2015 23:43:49 +0000 (15:43 -0800)]
Back-merge fixes + docs

and other fixes and documentation updates.

9 years agoMerge pull request #1756 from jeffdonahue/max-total-bytes-limit
Jon Long [Tue, 20 Jan 2015 00:59:48 +0000 (16:59 -0800)]
Merge pull request #1756 from jeffdonahue/max-total-bytes-limit

Max out the protobuf file read size limit

9 years agoSetTotalBytesLimit to the max (2 GB minus 1 byte)
Jeff Donahue [Thu, 16 Oct 2014 20:24:39 +0000 (13:24 -0700)]
SetTotalBytesLimit to the max (2 GB minus 1 byte)

9 years agoMerge pull request #1747 from yosinski/doc-up
Jon Long [Mon, 19 Jan 2015 04:29:47 +0000 (20:29 -0800)]
Merge pull request #1747 from yosinski/doc-up

Updated doc to suggest boost 1.57

9 years agoUpdated doc to suggest boost 1.57
Jason Yosinski [Sun, 18 Jan 2015 04:07:36 +0000 (23:07 -0500)]
Updated doc to suggest boost 1.57

9 years agoMerge pull request #1434 from pcampr/patch-1
Jeff Donahue [Sat, 17 Jan 2015 23:59:11 +0000 (15:59 -0800)]
Merge pull request #1434 from pcampr/patch-1

fixed filename in build_docs.sh

9 years agoMake comments for sparse GaussianFiller match actual behavior
Jeff Donahue [Sat, 17 Jan 2015 23:41:44 +0000 (15:41 -0800)]
Make comments for sparse GaussianFiller match actual behavior

(Fixes #1497 reported by @denizyuret)

9 years agoUpdate interfaces.md file
Christos Nikolaou [Wed, 29 Oct 2014 23:05:12 +0000 (01:05 +0200)]
Update interfaces.md file

Proofread and update the /docs/tutorial/interfaces.md file.

9 years agoMerge pull request #1388 from rohitgirdhar/cifar_docu_bug
Evan Shelhamer [Fri, 16 Jan 2015 23:59:50 +0000 (15:59 -0800)]
Merge pull request #1388 from rohitgirdhar/cifar_docu_bug

[docs] run CIFAR10 example from caffe root

9 years agoMerge pull request #1704 from longjon/friendlier-link-messages
Evan Shelhamer [Fri, 16 Jan 2015 22:41:52 +0000 (14:41 -0800)]
Merge pull request #1704 from longjon/friendlier-link-messages

Makefile: friendlier messages for link commands

9 years ago[docs] OpenCV version >= 2.4
Evan Shelhamer [Fri, 16 Jan 2015 21:41:34 +0000 (13:41 -0800)]
[docs] OpenCV version >= 2.4

9 years agoMerge pull request #1705 from longjon/origin-rpath
Evan Shelhamer [Fri, 16 Jan 2015 21:19:31 +0000 (13:19 -0800)]
Merge pull request #1705 from longjon/origin-rpath

Makefile: specify RPATH using $ORIGIN

9 years agoMerge pull request #1686 from longjon/net-const
Jon Long [Fri, 16 Jan 2015 19:21:49 +0000 (11:21 -0800)]
Merge pull request #1686 from longjon/net-const

Improve const-ness of Net

9 years agoMerge pull request #1662 from seanbell/fix-python-resize_image
Evan Shelhamer [Fri, 16 Jan 2015 06:31:25 +0000 (22:31 -0800)]
Merge pull request #1662 from seanbell/fix-python-resize_image

Fix caffe.io.resize_image for the case of constant images

9 years agoMerge pull request #1728 from shelhamer/pycaffe-mode-phase-device
Evan Shelhamer [Fri, 16 Jan 2015 05:59:16 +0000 (21:59 -0800)]
Merge pull request #1728 from shelhamer/pycaffe-mode-phase-device

Change Python interface for mode, phase, and device

9 years agocheck for enough args to convert_imageset
Evan Shelhamer [Fri, 16 Jan 2015 05:47:50 +0000 (21:47 -0800)]
check for enough args to convert_imageset

(this might better be handled by making all args flags...)

9 years agoMerge pull request #1236 from mlapin/legacy_nvcc_support
Evan Shelhamer [Fri, 16 Jan 2015 05:04:43 +0000 (21:04 -0800)]
Merge pull request #1236 from mlapin/legacy_nvcc_support

Drop OpenCV includes from NVCC code for legacy reasons.

9 years agoMerge pull request #1740 from shelhamer/yosemite-makefile
Evan Shelhamer [Fri, 16 Jan 2015 04:49:50 +0000 (20:49 -0800)]
Merge pull request #1740 from shelhamer/yosemite-makefile

Support OS X Yosemite / 10.10

9 years agolint internal thread
Evan Shelhamer [Fri, 16 Jan 2015 04:43:24 +0000 (20:43 -0800)]
lint internal thread

9 years agoMerge pull request #1335 from ryotat/master
Evan Shelhamer [Fri, 16 Jan 2015 04:43:53 +0000 (20:43 -0800)]
Merge pull request #1335 from ryotat/master

Fix leaking thread and groom internal thread implementation.

9 years agosupport OS X Yosemite / 10.10
Evan Shelhamer [Fri, 16 Jan 2015 00:28:09 +0000 (16:28 -0800)]
support OS X Yosemite / 10.10

- pick libstdc++ for OS X (regardless of version)
- make gtest rely on its own tuple to not conflict with clang
  (thanks @pluskid!)
- 10.10 has Accelerate while 10.9 has vecLib for BLAS
  (thanks @leonardt and @drdan14)

9 years agoset mode, phase, device in pycaffe; fix #1700
Evan Shelhamer [Thu, 15 Jan 2015 01:56:23 +0000 (17:56 -0800)]
set mode, phase, device in pycaffe; fix #1700

Attach mode, phase, and device setters to caffe module itself
so that these can be set before making nets. This is needed to properly
initialize layers with the right device and phase configuration.

Update examples to new usage.

9 years agoMerge pull request #1724 from pannous/wtf
Evan Shelhamer [Wed, 14 Jan 2015 19:41:37 +0000 (11:41 -0800)]
Merge pull request #1724 from pannous/wtf

Tell users to go to the caffe-users mailing list

9 years agoMessage: Please ask usage questions and how to model different tasks on the caffe...
Pannous [Wed, 14 Jan 2015 17:43:49 +0000 (18:43 +0100)]
Message: Please ask usage questions and how to model different tasks on the caffe-users mailing list

9 years ago[build] specify RPATH using $ORIGIN
Jonathan L Long [Sat, 10 Jan 2015 09:42:23 +0000 (01:42 -0800)]
[build] specify RPATH using $ORIGIN

Currently, when dynamically linking against libcaffe (right now, only
done for tests), RPATH is specified relative to the caffe source root.
This commit fixes RPATH using the special $ORIGIN variable to be
relative to the executable itself, so that there is no dependence on the
working directory.

9 years ago[build] more meaningful messages for link commands
Jonathan L Long [Sat, 10 Jan 2015 09:18:39 +0000 (01:18 -0800)]
[build] more meaningful messages for link commands

9 years agofix typo in layer_factory.cpp
Jonathan L Long [Sat, 10 Jan 2015 09:16:12 +0000 (01:16 -0800)]
fix typo in layer_factory.cpp

9 years agoimprove const-ness of Net
Jonathan L Long [Wed, 7 Jan 2015 04:25:53 +0000 (20:25 -0800)]
improve const-ness of Net

9 years agoBVLC models are for unrestricted use (follow-up to #1650)
Evan Shelhamer [Thu, 8 Jan 2015 04:00:27 +0000 (23:00 -0500)]
BVLC models are for unrestricted use (follow-up to #1650)

9 years ago[pycaffe] basic, partial testing of Net and SGDSolver
Jonathan L Long [Tue, 25 Nov 2014 02:53:14 +0000 (18:53 -0800)]
[pycaffe] basic, partial testing of Net and SGDSolver

9 years ago[travis] run pytest
Jonathan L Long [Fri, 21 Nov 2014 04:40:56 +0000 (20:40 -0800)]
[travis] run pytest

9 years ago[travis] remove unneeded Makefile.config sed hacking
Jonathan L Long [Tue, 25 Nov 2014 19:07:52 +0000 (11:07 -0800)]
[travis] remove unneeded Makefile.config sed hacking

Make allows variable redefinition, so sed brittleness can be avoided.

9 years agoadd "make pytest" for running Python tests
Jonathan L Long [Fri, 21 Nov 2014 04:36:56 +0000 (20:36 -0800)]
add "make pytest" for running Python tests

9 years agoMerge pull request #1228 from longjon/solver-step
Jeff Donahue [Wed, 7 Jan 2015 01:04:49 +0000 (17:04 -0800)]
Merge pull request #1228 from longjon/solver-step

Refactor Solver to allow interactive stepping

9 years ago[pycaffe] expose SGDSolver.step
Jonathan L Long [Mon, 29 Sep 2014 23:49:49 +0000 (16:49 -0700)]
[pycaffe] expose SGDSolver.step

9 years agobreak out Step from Solver
Jonathan L Long [Mon, 29 Sep 2014 23:47:06 +0000 (16:47 -0700)]
break out Step from Solver

9 years agoMerge pull request #1650 from shelhamer/unrestricted-bvlc-models
Evan Shelhamer [Fri, 2 Jan 2015 18:08:47 +0000 (13:08 -0500)]
Merge pull request #1650 from shelhamer/unrestricted-bvlc-models

Open release of BVLC model weights

9 years agofixed resize_image for the case of constant images
Sean Bell [Wed, 31 Dec 2014 01:06:31 +0000 (20:06 -0500)]
fixed resize_image for the case of constant images

9 years agoMerge pull request #1661 from longjon/init-test-labels
Jeff Donahue [Wed, 31 Dec 2014 00:02:31 +0000 (16:02 -0800)]
Merge pull request #1661 from longjon/init-test-labels

Fix sketchy label initialization in NetTest

9 years ago[tests] don't use Gaussian labels in NetTest's TinyNet
Jonathan L Long [Tue, 30 Dec 2014 20:52:53 +0000 (12:52 -0800)]
[tests] don't use Gaussian labels in NetTest's TinyNet

Previously, labels were being nonsensically filled with a zero-mean
Gaussian. Now they are just set to constant zero. Note that this is not
necessary for TrickyNet, neither forward nor backward are ever called.

9 years agoMerge pull request #1658 from jeffdonahue/make-everything-tweaks
Jon Long [Tue, 30 Dec 2014 10:33:43 +0000 (02:33 -0800)]
Merge pull request #1658 from jeffdonahue/make-everything-tweaks

'make everything' tweaks

9 years agodon't do runtest as part of 'make everything'
Jeff Donahue [Tue, 30 Dec 2014 09:33:05 +0000 (01:33 -0800)]
don't do runtest as part of 'make everything'

9 years agoonly build matcaffe as part of 'make everything' if MATLAB_DIR is set
Jeff Donahue [Tue, 30 Dec 2014 09:31:46 +0000 (01:31 -0800)]
only build matcaffe as part of 'make everything' if MATLAB_DIR is set

9 years agoMerge pull request #1656 from longjon/fix-softmax-loss-stubs
Jon Long [Tue, 30 Dec 2014 04:49:14 +0000 (20:49 -0800)]
Merge pull request #1656 from longjon/fix-softmax-loss-stubs

Fix an error made in removing the SoftmaxLossLayer GPU stubs

9 years agoMerge pull request #1449 from mprat/patch-1
Evan Shelhamer [Tue, 30 Dec 2014 03:39:07 +0000 (22:39 -0500)]
Merge pull request #1449 from mprat/patch-1

add pyyaml requirement for python

9 years agoremove SoftmaxLossLayer CPU_ONLY stubs, since there is no GPU version
Jonathan L Long [Tue, 30 Dec 2014 01:59:07 +0000 (17:59 -0800)]
remove SoftmaxLossLayer CPU_ONLY stubs, since there is no GPU version

9 years agoMerge pull request #1655 from longjon/softmax-loss-todo
Evan Shelhamer [Tue, 30 Dec 2014 01:48:07 +0000 (20:48 -0500)]
Merge pull request #1655 from longjon/softmax-loss-todo

Move softmax loss GPU todo comment from .cu file to header file

9 years agomove softmax loss GPU todo comment to header file
Jonathan L Long [Tue, 30 Dec 2014 01:36:08 +0000 (17:36 -0800)]
move softmax loss GPU todo comment to header file

There's no need to have GPU stubs just for a comment, and the absence of
a GPU-specific implementation is made more obvious by the nonexistence
of the .cu file.

9 years agoMerge pull request #1648 from longjon/find-warnings
Jon Long [Mon, 29 Dec 2014 22:57:41 +0000 (14:57 -0800)]
Merge pull request #1648 from longjon/find-warnings

Fix find warnings in upload_model_to_gist.sh

9 years ago[docs] open release of BVLC models for unrestricted use
Evan Shelhamer [Mon, 29 Dec 2014 21:23:18 +0000 (16:23 -0500)]
[docs] open release of BVLC models for unrestricted use

See BVLC model license details on the model zoo page.

9 years ago[docs] groom model zoo intro + list
Evan Shelhamer [Mon, 29 Dec 2014 20:13:59 +0000 (15:13 -0500)]
[docs] groom model zoo intro + list

9 years agoMerge pull request #1561 from longjon/pretty-build
Jon Long [Mon, 29 Dec 2014 05:39:27 +0000 (21:39 -0800)]
Merge pull request #1561 from longjon/pretty-build

Pretty up the build

9 years ago[scripts] fix find warnings in upload_model_to_gist.sh
Jonathan L Long [Mon, 29 Dec 2014 04:57:37 +0000 (20:57 -0800)]
[scripts] fix find warnings in upload_model_to_gist.sh

9 years agoBetter instructions for updating Homebrew after modifying formulae
Daniel Golden [Tue, 2 Dec 2014 19:20:58 +0000 (11:20 -0800)]
Better instructions for updating Homebrew after modifying formulae

Previous instructions glossed over how to deal with OpenCV-related update problems

9 years agoMerge pull request #1612 from sguada/googlenet_master
Evan Shelhamer [Mon, 29 Dec 2014 04:41:07 +0000 (23:41 -0500)]
Merge pull request #1612 from sguada/googlenet_master

BVLC GoogLeNet

9 years agoMerge pull request #1645 from longjon/remove-get-layer
Jeff Donahue [Mon, 29 Dec 2014 02:36:19 +0000 (18:36 -0800)]
Merge pull request #1645 from longjon/remove-get-layer

Remove unused GetLayer function (replaced by LayerRegistry::CreateLayer)

9 years agoremove unused GetLayer function (replaced by LayerRegistry::CreateLayer)
Jonathan L Long [Mon, 29 Dec 2014 02:03:45 +0000 (18:03 -0800)]
remove unused GetLayer function (replaced by LayerRegistry::CreateLayer)

9 years agopretty the build with the Q variable
Jonathan L Long [Thu, 11 Dec 2014 22:45:38 +0000 (14:45 -0800)]
pretty the build with the Q variable

Unset Q (in the environment) to see full commands, or edit in
Makefile.config.

9 years agoMerge pull request #1472 from longjon/incremental-build
Jon Long [Mon, 29 Dec 2014 01:36:24 +0000 (17:36 -0800)]
Merge pull request #1472 from longjon/incremental-build

Makefile improvements: rule consolidation, dependency generation

9 years agoautomatic dependency generation
Jonathan L Long [Mon, 24 Nov 2014 21:37:06 +0000 (13:37 -0800)]
automatic dependency generation

9 years agoMerge pull request #1636 from longjon/get-layer-gone
Yangqing Jia [Sun, 28 Dec 2014 21:34:43 +0000 (13:34 -0800)]
Merge pull request #1636 from longjon/get-layer-gone

Holiday merge - Update use of GetLayer to new registry function

9 years agoupdate use of GetLayer -> LayerRegistry::CreateLayer
Jonathan L Long [Sat, 27 Dec 2014 09:47:59 +0000 (01:47 -0800)]
update use of GetLayer -> LayerRegistry::CreateLayer

9 years agoMerge pull request #1585 from longjon/check-malloc
Sergio Guadarrama [Fri, 26 Dec 2014 19:13:40 +0000 (11:13 -0800)]
Merge pull request #1585 from longjon/check-malloc

Check host malloc result

9 years agoWarning of fallback only the first time for cudnn_pooling_layer
Sergio [Sun, 21 Dec 2014 22:03:49 +0000 (14:03 -0800)]
Warning of fallback only the first time for cudnn_pooling_layer

9 years agoMerge pull request #1607 from kjkjava/master
Jon Long [Mon, 22 Dec 2014 07:27:09 +0000 (23:27 -0800)]
Merge pull request #1607 from kjkjava/master

Use valid MathJax delimiters.

9 years agoAdded credits and bvlc_googlenet to model_zoo.md
Sergio [Sun, 21 Dec 2014 07:28:19 +0000 (23:28 -0800)]
Added credits and bvlc_googlenet to model_zoo.md

9 years agoAdded credits for training bvlc models
Sergio [Sun, 21 Dec 2014 07:24:32 +0000 (23:24 -0800)]
Added credits for training bvlc models

9 years agoAdded bvlc_googlenet prototxt and weights
Sergio [Tue, 2 Dec 2014 21:10:57 +0000 (13:10 -0800)]
Added bvlc_googlenet prototxt and weights

9 years agoFix init current_step
Sergio [Thu, 16 Oct 2014 18:17:24 +0000 (11:17 -0700)]
Fix init current_step

9 years agoAdapt lenet_multistep_solver.prototxt to current solvers
Sergio [Sat, 4 Oct 2014 00:14:20 +0000 (17:14 -0700)]
Adapt lenet_multistep_solver.prototxt to current solvers

9 years agoAdded Multistep, Poly and Sigmoid learning rate decay policies
Sergio [Sat, 4 Oct 2014 00:14:20 +0000 (17:14 -0700)]
Added Multistep, Poly and Sigmoid learning rate decay policies

Conflicts:
include/caffe/solver.hpp
src/caffe/proto/caffe.proto
src/caffe/solver.cpp

9 years agoDisplay averaged loss over the last several iterations
qipeng [Wed, 17 Sep 2014 02:59:53 +0000 (19:59 -0700)]
Display averaged loss over the last several iterations

9 years agoMerge pull request #1608 from sguada/added_credits
Evan Shelhamer [Sun, 21 Dec 2014 08:43:26 +0000 (00:43 -0800)]
Merge pull request #1608 from sguada/added_credits

Add credits for training BVLC Models

9 years agoAdded credits and bvlc_googlenet to model_zoo.md
Sergio [Sun, 21 Dec 2014 07:28:19 +0000 (23:28 -0800)]
Added credits and bvlc_googlenet to model_zoo.md

9 years agoAdded credits for training bvlc models
Sergio [Sun, 21 Dec 2014 07:24:32 +0000 (23:24 -0800)]
Added credits for training bvlc models

9 years agoMerge pull request #1598 from sguada/bvlc_googlenet
Sergio Guadarrama [Sun, 21 Dec 2014 05:13:40 +0000 (21:13 -0800)]
Merge pull request #1598 from sguada/bvlc_googlenet

Added bvlc_googlenet prototxt and weights

9 years agoAdded bvlc_googlenet prototxt and weights
Sergio [Tue, 2 Dec 2014 21:10:57 +0000 (13:10 -0800)]
Added bvlc_googlenet prototxt and weights

9 years agoFix minor typos in strings.
Kyle Krafka [Sun, 21 Dec 2014 04:34:51 +0000 (23:34 -0500)]
Fix minor typos in strings.

9 years agoUse valid MathJax delimiters.
Kyle Krafka [Sun, 21 Dec 2014 04:02:19 +0000 (23:02 -0500)]
Use valid MathJax delimiters.

MathJax does not support $ as a delimiter by default, so math was not
displayed properly on the “Forward and Backward” page.

9 years agoMerge pull request #1551 from schenker/fix_image_data_layer_segfault
Jeff Donahue [Sat, 20 Dec 2014 01:57:18 +0000 (17:57 -0800)]
Merge pull request #1551 from schenker/fix_image_data_layer_segfault

Add CHECKs to prevent segfault for incorrect IMAGE_DATA layers.

9 years agoAdd CHECKs to prevent segfault for incorrect IMAGE_DATA layers.
Thomas Schenker [Wed, 10 Dec 2014 15:16:34 +0000 (16:16 +0100)]
Add CHECKs to prevent segfault for incorrect IMAGE_DATA layers.

Incomplete IMAGE_DATA layers cause a segmentation fault, e.g.:

layers {
  name: "mnist"
  type: IMAGE_DATA
  top: "data"
  top: "label"
  image_data_param {
    batch_size: 100
  }
  include: { phase: TRAIN }
}

or

layers {
  name: "mnist"
  type: IMAGE_DATA
  top: "data"
  top: "label"
  data_param {
    source: "./data/train.txt"
    batch_size: 100
  }
  include: { phase: TRAIN }
}

Also empty image list files cause a segfault.
This is fixed by adding 3 CHECK instructions.

9 years agocheck host malloc result
Jonathan L Long [Wed, 17 Dec 2014 04:18:53 +0000 (20:18 -0800)]
check host malloc result

9 years agoMerge pull request #1558 from hojonathanho/hdf5_error_msg
Sergio Guadarrama [Tue, 16 Dec 2014 23:32:42 +0000 (15:32 -0800)]
Merge pull request #1558 from hojonathanho/hdf5_error_msg

Check input line count in HDF5 data layer

9 years agomove cuda output from build/.../.cuo -> build/cuda/.../.o
Jonathan L Long [Thu, 11 Dec 2014 07:37:42 +0000 (23:37 -0800)]
move cuda output from build/.../.cuo -> build/cuda/.../.o

This will allow nvcc's -M dependency generation option to work
harmoniously, since it assumes that output will have a .o extension.

9 years agoautomatic discovery of source directories
Jonathan L Long [Mon, 24 Nov 2014 08:10:35 +0000 (00:10 -0800)]
automatic discovery of source directories

9 years agoconsolidate build rules for object files
Jonathan L Long [Mon, 24 Nov 2014 06:35:46 +0000 (22:35 -0800)]
consolidate build rules for object files

9 years agoremove extra blank line
Jonathan L Long [Tue, 16 Dec 2014 09:52:43 +0000 (01:52 -0800)]
remove extra blank line

9 years agoCheck input line count in HDF5 data layer
Jonathan Ho [Thu, 11 Dec 2014 02:27:56 +0000 (18:27 -0800)]
Check input line count in HDF5 data layer

9 years agoMerge pull request #1384 from CellScope/log-parser-python
Evan Shelhamer [Mon, 8 Dec 2014 22:00:49 +0000 (17:00 -0500)]
Merge pull request #1384 from CellScope/log-parser-python

Python log parser

9 years agoStore data in lists of dicts and use csv package
Daniel Golden [Mon, 3 Nov 2014 05:02:07 +0000 (21:02 -0800)]
Store data in lists of dicts and use csv package

Output format is unchanged (except that csv.DictWriter insists on writing ints as 0.0 instead of 0)

9 years agoTake train loss from `Iteration N, loss = X` lines
Daniel Golden [Fri, 31 Oct 2014 17:24:07 +0000 (10:24 -0700)]
Take train loss from `Iteration N, loss = X` lines

Was previously using `Train net output #M: loss = X` lines, but there may not be exactly one of those (e.g., for GoogLeNet, which has multiple loss layers); I believe that `Iteration N, loss = X` is the aggregated loss.

If there's only one loss layer, these two values will be equal and it won't matter. Otherwise, we presumably want to report the aggregated loss.

9 years agoCreated parse_log.py, competitor to parse_log.sh
Daniel Golden [Thu, 30 Oct 2014 22:51:14 +0000 (15:51 -0700)]
Created parse_log.py, competitor to parse_log.sh

9 years agoMerge pull request #1406 from CellScope/matcaffe-osx-fix-crash-on-error
Evan Shelhamer [Mon, 8 Dec 2014 19:50:02 +0000 (14:50 -0500)]
Merge pull request #1406 from CellScope/matcaffe-osx-fix-crash-on-error

Prevent Matlab on OS X from crashing on error

9 years agoMerge pull request #1516 from drdan14/update-homebrew-install-instructions
Evan Shelhamer [Mon, 8 Dec 2014 19:45:50 +0000 (14:45 -0500)]
Merge pull request #1516 from drdan14/update-homebrew-install-instructions

Better instructions for updating Homebrew after modifying formulae

9 years agoMerge pull request #1527 from drdan14/classify-py-mean-dims
Evan Shelhamer [Mon, 8 Dec 2014 19:44:15 +0000 (14:44 -0500)]
Merge pull request #1527 from drdan14/classify-py-mean-dims

Update mean file help

9 years agoclarify #endif comment
Jonathan L Long [Sun, 7 Dec 2014 11:06:35 +0000 (03:06 -0800)]
clarify #endif comment

9 years agocompile for compute capability 5.0
Evan Shelhamer [Fri, 5 Dec 2014 16:06:40 +0000 (08:06 -0800)]
compile for compute capability 5.0

9 years agoUpdate mean file help
Daniel Golden [Wed, 3 Dec 2014 22:30:14 +0000 (14:30 -0800)]
Update mean file help

Numpy array dimensions should be [Channels x Height x Width], NOT [Height x Width x Channels]

If an image mean file is supplied with channels as the last dimension, classify.py throws errors like:

ValueError: operands could not be broadcast together with shapes (3,227,227) (256,0,225) (3,227,227)