platform/upstream/caffeonacl.git
10 years agolink to demo
Sergey Karayev [Fri, 23 May 2014 01:18:25 +0000 (18:18 -0700)]
link to demo

10 years agoMerge pull request #435 from shelhamer/v1-models
Evan Shelhamer [Thu, 22 May 2014 08:04:36 +0000 (01:04 -0700)]
Merge pull request #435 from shelhamer/v1-models

Release v1 model defs + weights

10 years agofix draw_net python script
Evan Shelhamer [Thu, 22 May 2014 08:02:56 +0000 (01:02 -0700)]
fix draw_net python script

include caffe.draw for drawing functions.

10 years agorelease v1 model defs + weights
Evan Shelhamer [Thu, 22 May 2014 07:56:35 +0000 (00:56 -0700)]
release v1 model defs + weights

- Caffe reference ImageNet model
- AlexNet

Note that one can upgrade the weights locally by
`upgrade_net_proto_binary.bin` to avoid re-downloading.

10 years agopoint out @niuzhiheng's work on the Windows port
Evan Shelhamer [Thu, 22 May 2014 06:43:49 +0000 (23:43 -0700)]
point out @niuzhiheng's work on the Windows port

10 years agofix test_all path in docs
Evan Shelhamer [Wed, 21 May 2014 17:52:45 +0000 (10:52 -0700)]
fix test_all path in docs

10 years agolink canonical bvlc site
Evan Shelhamer [Tue, 20 May 2014 21:44:47 +0000 (14:44 -0700)]
link canonical bvlc site

10 years agofix detection notebook link
Evan Shelhamer [Tue, 20 May 2014 21:42:37 +0000 (14:42 -0700)]
fix detection notebook link

10 years agoMerge pull request #429 from shelhamer/next
Evan Shelhamer [Tue, 20 May 2014 21:20:15 +0000 (14:20 -0700)]
Merge pull request #429 from shelhamer/next

Next: 0.999

10 years agoBack-merge changes in master
Evan Shelhamer [Tue, 20 May 2014 19:44:51 +0000 (12:44 -0700)]
Back-merge changes in master

* master:
  bundle presentation in gh-pages for now...
  fix typo pointed out by @yinxusen
  note support for non-MKL installation in dev
  include pretrained snapshot and performance details
  Document AlexNet model, include download script
  define AlexNet architecture
  polished ignore

10 years agoMerge pull request #311 from shelhamer/python-fixes
Evan Shelhamer [Tue, 20 May 2014 19:20:00 +0000 (12:20 -0700)]
Merge pull request #311 from shelhamer/python-fixes

Improve python wrapper

10 years agoupdate notebook examples with new wrapper usage, re-organize
Evan Shelhamer [Tue, 20 May 2014 08:04:17 +0000 (01:04 -0700)]
update notebook examples with new wrapper usage, re-organize

10 years agopreprocess single inputs instead of lists
Evan Shelhamer [Tue, 20 May 2014 18:41:42 +0000 (11:41 -0700)]
preprocess single inputs instead of lists

For compositionality and expectations.

10 years agowindowed detection in python
Evan Shelhamer [Tue, 20 May 2014 06:50:15 +0000 (23:50 -0700)]
windowed detection in python

10 years agosquash infuriating loop assignment bug in batching
Evan Shelhamer [Tue, 20 May 2014 05:55:50 +0000 (22:55 -0700)]
squash infuriating loop assignment bug in batching

10 years agoimage classification in python
Evan Shelhamer [Mon, 19 May 2014 22:31:49 +0000 (15:31 -0700)]
image classification in python

10 years agofix padding for the last batch
Evan Shelhamer [Mon, 19 May 2014 01:25:18 +0000 (18:25 -0700)]
fix padding for the last batch

10 years agosplit drawnet into module code and script
Evan Shelhamer [Mon, 19 May 2014 00:14:53 +0000 (17:14 -0700)]
split drawnet into module code and script

Don't run scripts in the module dir to avoid import collisions between
io and caffe.io.

10 years agoadd caffe.io submodule for conversions, image loading and resizing
Evan Shelhamer [Mon, 19 May 2014 00:13:05 +0000 (17:13 -0700)]
add caffe.io submodule for conversions, image loading and resizing

10 years agofix python mean subtraction
Evan Shelhamer [Mon, 19 May 2014 00:11:38 +0000 (17:11 -0700)]
fix python mean subtraction

10 years agoMerge pull request #376 from sergeyk/layer_reorg
Sergey Karayev [Mon, 19 May 2014 22:50:33 +0000 (15:50 -0700)]
Merge pull request #376 from sergeyk/layer_reorg

Layer definitions and declarations re-organization and documentation

10 years agoIncorporated Evan’s comments for neuron layers
Sergey Karayev [Mon, 19 May 2014 18:11:37 +0000 (11:11 -0700)]
Incorporated Evan’s comments for neuron layers

10 years agoCosmetic change in ConcatLayer
Sergey Karayev [Mon, 19 May 2014 17:44:21 +0000 (10:44 -0700)]
Cosmetic change in ConcatLayer

10 years agoLil’ more docstring, and cosmetic change in EuclideanLossLayer
Sergey Karayev [Mon, 19 May 2014 17:43:21 +0000 (10:43 -0700)]
Lil’ more docstring, and cosmetic change in EuclideanLossLayer

10 years agofwd/back math docs for neuron layers
Sergey Karayev [Tue, 29 Apr 2014 07:21:15 +0000 (00:21 -0700)]
fwd/back math docs for neuron layers

10 years agodrop cute names in favor of Net.{pre,de}process() for input formatting
Evan Shelhamer [Fri, 16 May 2014 23:03:55 +0000 (16:03 -0700)]
drop cute names in favor of Net.{pre,de}process() for input formatting

...and refer to inputs as inputs and not images since general vectors
and matrices are perfectly fine.

10 years agoNet.caffeinate() and Net.decaffeinate() format/unformat lists
Evan Shelhamer [Fri, 16 May 2014 01:52:07 +0000 (18:52 -0700)]
Net.caffeinate() and Net.decaffeinate() format/unformat lists

10 years agotake blob args as ndarrays and assign on the python side
Evan Shelhamer [Fri, 16 May 2014 01:14:45 +0000 (18:14 -0700)]
take blob args as ndarrays and assign on the python side

Take blob args and give blob returns as single ndarrays instead of lists
of arrays.

Assign the net blobs and diffs as needed on the python side, which
reduces copies and simplifies the C++ side of the wrapper.

Thanks @longjon for the suggestion.

10 years agoCosmetic change in prep for data layer work
Sergey Karayev [Tue, 29 Apr 2014 02:40:43 +0000 (19:40 -0700)]
Cosmetic change in prep for data layer work

10 years agoSplit all loss layers into own .cpp files
Sergey Karayev [Tue, 29 Apr 2014 02:39:36 +0000 (19:39 -0700)]
Split all loss layers into own .cpp files

10 years agolayer definition reorganization and documentation
Sergey Karayev [Tue, 29 Apr 2014 02:06:07 +0000 (19:06 -0700)]
layer definition reorganization and documentation
- split out neuron, loss, and data layers into own header files
- added LossLayer class with common SetUp checks
- in-progress concise documentation of each layer's purpose

10 years agoresize to input dimensions when formatting in python
Evan Shelhamer [Thu, 15 May 2014 20:52:07 +0000 (13:52 -0700)]
resize to input dimensions when formatting in python

10 years agoreplace iterator with indices for consistency
Evan Shelhamer [Thu, 15 May 2014 20:06:17 +0000 (13:06 -0700)]
replace iterator with indices for consistency

10 years agopython style
Evan Shelhamer [Thu, 15 May 2014 19:35:32 +0000 (12:35 -0700)]
python style

10 years agofix accidental revert of Init() from f5c28581
Evan Shelhamer [Thu, 15 May 2014 17:20:43 +0000 (10:20 -0700)]
fix accidental revert of Init() from f5c28581

10 years agobatch inputs in python by forward_all() and forward_backward_all()
Evan Shelhamer [Thu, 15 May 2014 06:45:33 +0000 (23:45 -0700)]
batch inputs in python by forward_all() and forward_backward_all()

10 years agodon't squeeze blob arrays for python
Evan Shelhamer [Thu, 15 May 2014 03:17:09 +0000 (20:17 -0700)]
don't squeeze blob arrays for python

Preserve the non-batch dimensions of blob arrays, even for singletons.

The forward() and backward() helpers take lists of ndarrays instead of a
single ndarray per blob, and lists of ndarrays are likewise returned.

Note that for output the blob array could actually be returned as a
single ndarray instead of a list.

10 years agopython forward() and backward() extract any blobs and diffs
Evan Shelhamer [Thu, 15 May 2014 00:38:33 +0000 (17:38 -0700)]
python forward() and backward() extract any blobs and diffs

10 years agopython Net.backward() helper and Net.BackwardPrefilled()
Evan Shelhamer [Wed, 14 May 2014 21:37:33 +0000 (14:37 -0700)]
python Net.backward() helper and Net.BackwardPrefilled()

10 years agobad forward/backward inputs throw exceptions instead of crashing python
Evan Shelhamer [Wed, 14 May 2014 21:02:54 +0000 (14:02 -0700)]
bad forward/backward inputs throw exceptions instead of crashing python

10 years agopycaffe Net.forward() helper
Evan Shelhamer [Wed, 14 May 2014 20:39:06 +0000 (13:39 -0700)]
pycaffe Net.forward() helper

Do forward pass by prefilled or packaging input + output blobs and
returning a {output blob name: output list} dict.

10 years agoset input preprocessing per blob in python
Evan Shelhamer [Wed, 14 May 2014 02:56:14 +0000 (19:56 -0700)]
set input preprocessing per blob in python

10 years agoexpose input and output blob names to python as lists
Evan Shelhamer [Wed, 14 May 2014 01:53:36 +0000 (18:53 -0700)]
expose input and output blob names to python as lists

10 years agoMerge pull request #417 from shelhamer/create-and-write-proto
Jeff Donahue [Wed, 14 May 2014 20:05:17 +0000 (13:05 -0700)]
Merge pull request #417 from shelhamer/create-and-write-proto

Write/create/truncate prototxt when saving to fix #341

10 years agofix workaround in net prototxt upgrade
Evan Shelhamer [Wed, 14 May 2014 19:35:33 +0000 (12:35 -0700)]
fix workaround in net prototxt upgrade

10 years agoWrite/create/truncate prototxt when saving to fix #341
Evan Shelhamer [Wed, 14 May 2014 19:28:55 +0000 (12:28 -0700)]
Write/create/truncate prototxt when saving to fix #341

WriteProtoToTextFile now saves the prototxt whether or not the file
already exists and sets the permissions to owner read/write and group +
other read.

Thanks @beam2d and @chyh1990 for pointing out the open modes bug.

10 years agopycaffe comments, lint
Evan Shelhamer [Thu, 10 Apr 2014 03:27:58 +0000 (20:27 -0700)]
pycaffe comments, lint

10 years agoadd python io getters, mean helper, and image caffeinator/decaffeinator
Evan Shelhamer [Thu, 10 Apr 2014 01:37:32 +0000 (18:37 -0700)]
add python io getters, mean helper, and image caffeinator/decaffeinator

10 years agomake python wrapper mean match binaryproto dimensions
Evan Shelhamer [Thu, 10 Apr 2014 01:33:18 +0000 (18:33 -0700)]
make python wrapper mean match binaryproto dimensions

ilsvrc_2012_mean.npy has dims K x H x W.

Code written for the old D x D x K mean needs to be rewritten!

10 years agomatch existing python formatting
Evan Shelhamer [Wed, 9 Apr 2014 21:02:40 +0000 (14:02 -0700)]
match existing python formatting

10 years agoMerge pull request #414 from shelhamer/net-output-blobs
Evan Shelhamer [Wed, 14 May 2014 01:10:22 +0000 (18:10 -0700)]
Merge pull request #414 from shelhamer/net-output-blobs

Make net know output blob indices

10 years agonet knows output blobs
Evan Shelhamer [Wed, 14 May 2014 01:08:06 +0000 (18:08 -0700)]
net knows output blobs

10 years agoMerge pull request #413 from shelhamer/cublas-status-not-supported
Evan Shelhamer [Tue, 13 May 2014 20:16:19 +0000 (13:16 -0700)]
Merge pull request #413 from shelhamer/cublas-status-not-supported

add cublas status in cuda 6 to fix warning

10 years agoadd cublas status in cuda 6 to fix warning
Evan Shelhamer [Tue, 13 May 2014 19:30:34 +0000 (12:30 -0700)]
add cublas status in cuda 6 to fix warning

...and #define for older CUDAs to not break the build.

10 years agoMerge pull request #406 from jeffdonahue/makefile-include-bug
Jeff Donahue [Sun, 11 May 2014 00:41:39 +0000 (17:41 -0700)]
Merge pull request #406 from jeffdonahue/makefile-include-bug

Fix Makefile header dependency bug

10 years agofix Makefile bug - HXX_SRCS was things that don't end in .hpp, instead
Jeff Donahue [Sun, 11 May 2014 00:20:19 +0000 (17:20 -0700)]
fix Makefile bug - HXX_SRCS was things that don't end in .hpp, instead
of things that do...

10 years agoMerge pull request #403 from jeffdonahue/solver-mode-enum
Evan Shelhamer [Fri, 9 May 2014 21:56:13 +0000 (14:56 -0700)]
Merge pull request #403 from jeffdonahue/solver-mode-enum

Make solver_mode an enum: CPU or GPU

10 years agomake solver_mode an enum with CPU and GPU -- fully backwards compatible
Jeff Donahue [Fri, 9 May 2014 21:50:05 +0000 (14:50 -0700)]
make solver_mode an enum with CPU and GPU -- fully backwards compatible
with old 0/1 style

10 years agoMerge pull request #396 from longjon/math-includes
Jeff Donahue [Wed, 7 May 2014 04:34:50 +0000 (21:34 -0700)]
Merge pull request #396 from longjon/math-includes

Improve includes in util/math_functions.hpp

10 years agoimprove includes in util/math_function.hpp
Jonathan L Long [Tue, 6 May 2014 23:40:09 +0000 (16:40 -0700)]
improve includes in util/math_function.hpp

This commit removes the redundant <math.h>, and adds the necessary
<stdint.h> and <glog/logging.h>.

10 years agonote bug in cifar10 full with CPU computation
Jeff Donahue [Tue, 6 May 2014 18:59:25 +0000 (11:59 -0700)]
note bug in cifar10 full with CPU computation

10 years agobundle presentation in gh-pages for now...
Evan Shelhamer [Tue, 6 May 2014 17:04:13 +0000 (10:04 -0700)]
bundle presentation in gh-pages for now...

...dropbox disabled our link. Actual solution: configure web server for
models and other materials soon.

10 years agoMerge pull request #294 from longjon/memory-data-layer
Evan Shelhamer [Fri, 2 May 2014 21:11:31 +0000 (14:11 -0700)]
Merge pull request #294 from longjon/memory-data-layer

Add a layer for in-memory data, and expose it to Python

10 years agofix lint error in syncedmem.hpp
Jonathan L Long [Fri, 25 Apr 2014 21:56:44 +0000 (14:56 -0700)]
fix lint error in syncedmem.hpp

10 years agopycaffe: allow 1d labels to be passed to set_input_arrays
Jonathan L Long [Thu, 17 Apr 2014 10:06:20 +0000 (03:06 -0700)]
pycaffe: allow 1d labels to be passed to set_input_arrays

10 years agopycaffe: add Net.set_input_arrays for input from numpy
Jonathan L Long [Fri, 25 Apr 2014 21:33:25 +0000 (14:33 -0700)]
pycaffe: add Net.set_input_arrays for input from numpy

This requires a net whose first layer is a MemoryDataLayer.

10 years agopycaffe: store a shared_ptr<CaffeNet> in SGDSolver
Jonathan L Long [Fri, 25 Apr 2014 21:29:50 +0000 (14:29 -0700)]
pycaffe: store a shared_ptr<CaffeNet> in SGDSolver

Doing this, rather than constructing the CaffeNet wrapper every time,
will allow the wrapper to hold references that last at least as long as
SGDSolver (which will be necessary to ensure that data used by
MemoryDataLayer doesn't get freed).

10 years agopycaffe: let boost pass shared_ptr<CaffeNet>
Jonathan L Long [Fri, 25 Apr 2014 21:09:02 +0000 (14:09 -0700)]
pycaffe: let boost pass shared_ptr<CaffeNet>

10 years agoadd basic tests for MemoryDataLayer
Jonathan L Long [Thu, 17 Apr 2014 09:24:02 +0000 (02:24 -0700)]
add basic tests for MemoryDataLayer

10 years agoadd size accessors to MemoryDataLayer
Jonathan L Long [Fri, 4 Apr 2014 22:20:02 +0000 (15:20 -0700)]
add size accessors to MemoryDataLayer

This will facilitate input size checking for pycaffe (and potentially
others).

10 years agoadd MemoryDataLayer for reading input from contiguous blocks of memory
Jonathan L Long [Fri, 4 Apr 2014 02:27:21 +0000 (19:27 -0700)]
add MemoryDataLayer for reading input from contiguous blocks of memory

10 years agoadd set_cpu_data to Blob and SyncedMemory
Jonathan L Long [Fri, 4 Apr 2014 02:19:47 +0000 (19:19 -0700)]
add set_cpu_data to Blob and SyncedMemory

This allows a blob to be updated without copy to use already existing
memory (and will support MemoryDataLayer).

10 years agoMerge pull request #378 from longjon/no-merge-duplicates
Jeff Donahue [Wed, 30 Apr 2014 19:53:32 +0000 (12:53 -0700)]
Merge pull request #378 from longjon/no-merge-duplicates

Note the last added layer/params in caffe.proto to prevent conflicts

10 years agonote the last added layer/params in caffe.proto to prevent conflicts
Jonathan L Long [Wed, 30 Apr 2014 19:28:55 +0000 (12:28 -0700)]
note the last added layer/params in caffe.proto to prevent conflicts

The current scheme does not actually prevent conflicts, since three-way
merge will accept simultaneous changes that agree on the next number.
This commit fixes this by explicitly noting the last layer added.

10 years agoMerge pull request #377 from sguada/fix_initial_test
Evan Shelhamer [Wed, 30 Apr 2014 18:37:56 +0000 (11:37 -0700)]
Merge pull request #377 from sguada/fix_initial_test

Keep same format for all net testing log messages

10 years agoKeep uniform test messages for all the test
Sergio Guadarrama [Wed, 30 Apr 2014 18:26:55 +0000 (11:26 -0700)]
Keep uniform test messages for all the test

10 years agorollback 8368818, does not build
Jeff Donahue [Tue, 29 Apr 2014 04:50:00 +0000 (21:50 -0700)]
rollback 8368818, does not build

10 years agoAdd Sublime Text project settings to gitignore
Sergey Karayev [Tue, 29 Apr 2014 02:47:17 +0000 (19:47 -0700)]
Add Sublime Text project settings to gitignore

10 years agoHandling CUBLAS_STATUS_NOT_SUPPORTED to suppress warning
Sergey Karayev [Tue, 29 Apr 2014 02:44:32 +0000 (19:44 -0700)]
Handling CUBLAS_STATUS_NOT_SUPPORTED to suppress warning

10 years agoMerge pull request #370 from shelhamer/test-net-polish
Jeff Donahue [Sat, 26 Apr 2014 23:46:39 +0000 (16:46 -0700)]
Merge pull request #370 from shelhamer/test-net-polish

Polish test_net

10 years agofix test_net to upgrade params from v0 if needed
Evan Shelhamer [Sat, 26 Apr 2014 23:33:40 +0000 (16:33 -0700)]
fix test_net to upgrade params from v0 if needed

10 years agodefault test net device to 0 and log device chosen
Evan Shelhamer [Sat, 26 Apr 2014 23:08:22 +0000 (16:08 -0700)]
default test net device to 0 and log device chosen

10 years agoset seed in neuron and power layer tests for deterministic results
Jeff Donahue [Sat, 26 Apr 2014 22:47:35 +0000 (15:47 -0700)]
set seed in neuron and power layer tests for deterministic results

10 years agoMerge pull request #367 from shelhamer/randomize-test-order
Jeff Donahue [Sat, 26 Apr 2014 22:45:11 +0000 (15:45 -0700)]
Merge pull request #367 from shelhamer/randomize-test-order

Randomize order of test execution by `make runtest`

10 years agoproofreading and trivial polish
Evan Shelhamer [Sat, 26 Apr 2014 22:23:56 +0000 (15:23 -0700)]
proofreading and trivial polish

10 years agoadd device id arg to test_net (fix #232)
Evan Shelhamer [Sat, 26 Apr 2014 22:22:42 +0000 (15:22 -0700)]
add device id arg to test_net (fix #232)

10 years agoMerge pull request #303 from longjon/hinge-loss-layer
Jeff Donahue [Sat, 26 Apr 2014 22:13:15 +0000 (15:13 -0700)]
Merge pull request #303 from longjon/hinge-loss-layer

HingeLossLayer

10 years agoMerge pull request #368 from shelhamer/fix-data-layer-sequence-tests
Jeff Donahue [Sat, 26 Apr 2014 22:09:17 +0000 (15:09 -0700)]
Merge pull request #368 from shelhamer/fix-data-layer-sequence-tests

Fix data layer sequence tests to avoid leveldb contention and lock failures

10 years agoscope data layer sequence tests to avoid leveldb contention
Evan Shelhamer [Sat, 26 Apr 2014 21:57:11 +0000 (14:57 -0700)]
scope data layer sequence tests to avoid leveldb contention

10 years agoset phase in dropout tests
Evan Shelhamer [Sat, 26 Apr 2014 21:42:38 +0000 (14:42 -0700)]
set phase in dropout tests

10 years agorandomize order of test execution by make runtest
Evan Shelhamer [Sat, 26 Apr 2014 01:54:12 +0000 (18:54 -0700)]
randomize order of test execution by make runtest

10 years agoMerge pull request #366 from shelhamer/test-phase-fix
Jeff Donahue [Sat, 26 Apr 2014 01:42:37 +0000 (18:42 -0700)]
Merge pull request #366 from shelhamer/test-phase-fix

set TRAIN in CommonTest.TestPhase

10 years agoset TRAIN in CommonTest.TestPhase
Evan Shelhamer [Sat, 26 Apr 2014 01:33:40 +0000 (18:33 -0700)]
set TRAIN in CommonTest.TestPhase

...to avoid effects of test execution order. The original intention was
to check that TRAIN is the default, but it is better to have consistent
results.

10 years agotest HingeLossLayer
Jonathan L Long [Fri, 25 Apr 2014 23:44:05 +0000 (16:44 -0700)]
test HingeLossLayer

Based on SoftmaxWithLossLayerTest.

10 years agomake gradient checker's kink use feature absolute value
Jonathan L Long [Fri, 25 Apr 2014 23:38:48 +0000 (16:38 -0700)]
make gradient checker's kink use feature absolute value

In theory, layer functions could be nonsmooth anywhere; in all cases in
use so far, they are nonsmooth at either zero or +1 and -1. In the
future, it might be necessary to generalize the kink mechanism beyond
this stopgap measure.

10 years agoadd HingeLossLayer for one-vs-all hinge loss
Jonathan L Long [Tue, 8 Apr 2014 03:35:33 +0000 (20:35 -0700)]
add HingeLossLayer for one-vs-all hinge loss

This layer implements a "one-vs-all" hinge loss, (1/n) sum_ij max(0, 1 -
y_ij x_ij), with bottom blob x_ij (i ranging over examples and j over
classes), and y_ij = +1/-1 indicating the label. No regularization is
included, since regularization is done via weight decay or using the
parameters of another layer. The gradient is taken to be zero at the
hinge point. This commit only provides the CPU implementation.

10 years agoMerge pull request #365 from longjon/pycaffe-empty-nets
Evan Shelhamer [Fri, 25 Apr 2014 23:02:37 +0000 (16:02 -0700)]
Merge pull request #365 from longjon/pycaffe-empty-nets

Add unary CaffeNet constructor for uninitialized nets

10 years agopycaffe: add unary CaffeNet constructor for uninitialized nets
Jonathan L Long [Fri, 25 Apr 2014 21:50:36 +0000 (14:50 -0700)]
pycaffe: add unary CaffeNet constructor for uninitialized nets

Note that parameters are uninitialized, not zero-filled.

10 years agoMerge pull request #363 from jeffdonahue/speedup-gradient-check
Evan Shelhamer [Fri, 25 Apr 2014 19:51:22 +0000 (12:51 -0700)]
Merge pull request #363 from jeffdonahue/speedup-gradient-check

Gradient check speedups

10 years agoeltwise gradient checker
Jeff Donahue [Fri, 25 Apr 2014 03:24:19 +0000 (20:24 -0700)]
eltwise gradient checker