32ecc191f1
Merge pull request #2100 from Yancey1989/fix2060
...
Bugfix: checkout input_type failed
8 years ago
3c5e0f8a5d
remove g_pass_height_width in config_parse.py
8 years ago
85502783aa
Bugfix: checkout input_type failed
8 years ago
1fd00758ac
SmoothL1 loss
8 years ago
d369577fed
add reversed poolSequenceWithStride
8 years ago
0291c01884
Merge branch 'develop' into stride
8 years ago
5961b52b13
Merge pull request #1653 from Noplz/normalize-layer
...
CrossChannelNorm Layer for SSD
8 years ago
9298a9ec0d
stride pooling for seqlastin and seqfirstin
8 years ago
ec7184ac18
fixed a bug in conv_projection issue #1681
8 years ago
090c974e4b
completed implementation of cudnn_convt convTransProjection and convTransOperator
8 years ago
07c1ea258f
python interface for convTransProjection and convTransOperator
8 years ago
8c2d1bad7a
Merge branch 'develop' into normalize-layer
8 years ago
eb43d93a58
Change Normalize layer to CrossChannelNorm layer
8 years ago
bd4ec1b493
Merge branch 'develop' into memory.set_input
8 years ago
03889bab6f
Merge branch 'develop' into fix-crf-weight-and-coeff-bug
8 years ago
4375a64ea1
fix bug to pass travis
8 years ago
bf838034d6
Merge branch 'develop' into fix-crf-weight-and-coeff-bug
8 years ago
eea0097dcb
NormalizeLayer for SSD
8 years ago
1d58949a40
Merge branch 'develop' into enable_drop_in_average_and_max_layer
8 years ago
fd41a87a53
complete data_type documentation
8 years ago
e87181db8d
use value_range instead of ele_range
8 years ago
fa3b1c465b
change argument name of data_type.integer_value/integer_value_sequence to ele_range, add documentation.
8 years ago
e768721cd9
fix calculating totalScore2_ bug
8 years ago
7d551dd484
Make it possible to postpone setting the layer name for a memory.
...
The reason for adding the function is exemplified in the following hypothetical code:
mem = memory(name=None, size=256)
hidden = fc_layer(input=mem)
state = hidden + x
mem.set_input(state)
In the above code segment, it would be very annoying if we require the user to provide
the name at memory() call because the layer name of state is automatically generated
and is not easy to set it.
Change-Id: I918bf1d3d5c26addd88a6f7021e98b3e0e9df494
8 years ago
ccb553fec4
Merge pull request #1253 from wangkuiyi/python_learning_and_refactor
...
Rename Python function DataBase into create_data_config_proto
8 years ago
b8f3a5c433
Correct the wrong deletion of a single character
8 years ago
b36df98d01
Correct Python identation
8 years ago
058eeac0fc
Revert "Remove completely create_data_config_proto"
...
This reverts commit ab279beed1
8 years ago
ab279beed1
Remove completely create_data_config_proto
8 years ago
996b1de1d8
Rename DataBase into create_data_config_proto
8 years ago
6245fed240
rotate_layer python interface fixes
8 years ago
55eb2fcffa
format correction
8 years ago
781b85b5fc
rotate_layer and flip_layer * added getMin and getMax for GpuMatrix * gru_step_layer parameter name
8 years ago
04b5daf92d
change the parameter position of gru_step_layer from 1 back to 0
8 years ago
6da7283475
make gru_group parameters sharable
8 years ago
8d47499e1d
update code
8 years ago
19cb2a7fb3
fix default value of cos_sim scale
8 years ago
6b61a096e1
Optional padding mode, namely ceil or floor, ceil by default.
8 years ago
8b5431d5e1
padding operation
8 years ago
cbbad4202a
delete the modification of SlopeInterceptLayer.
8 years ago
08a817e3fb
delete unnecessary parameters and modifications for some mathmatical
...
layers.
8 years ago
ce939b30ba
enable dropout rate in several computation layers.
8 years ago
15a6252f8e
Merge branch 'develop' into enable_drop_in_average_and_max_layer
8 years ago
e022c065d7
enable dropout in average and max layer.
8 years ago
843b63bb84
add config_parser in trainer_config_helpers to seperate trainer config
8 years ago
c6d9d04004
Merge pull request #979 from reyoung/feature/fix_param_hidden_in_pydp2
...
A tiny fix in PyDataProvider2
8 years ago
16ea66e8ed
Merge branch 'develop' of github.com:baidu/Paddle into feature/mnist_train_api
8 years ago
e031f0c4e8
Fix typo in PyDataProvider2.py
8 years ago
b53bdcdc67
Merge pull request #867 from Noplz/ssd
...
priorbox layer for Single Shot Multibox Detection Network
8 years ago
b063b0c405
Merge branch 'develop' into fix-crf-weight-and-coeff-bug
8 years ago
4d81b36123
A tiny fix in PyDataProvider2
...
* hidden decorator kwargs in DataProvider.__init__
* also add unit test for this.
8 years ago
af820fb2eb
Merge branch 'develop' into fix-crf-weight-and-coeff-bug
8 years ago
e4c492d3b8
change type to bool.
8 years ago
5bb29ece7f
close log info in BN.
8 years ago
f4f0f2daeb
Fix bug in config_parse.py when batch_norm layer is used in RecurrentLayerGroup.
8 years ago
f1e31ecada
Merge branch 'develop' into fix-crf-weight-and-coeff-bug
8 years ago
8fe3a3aa73
Add excluded_chunk_types to ChunkEvaluator
...
The chunks of types in excluded_chunk_types will not be counted in
ChunkEvaluator. This is useful for tasks such as SRL, in which chunks of
type V (verb) will not be taken into account in evaluation.
8 years ago
50edf5243e
Merge pull request #862 from backyes/fix_data_sources
...
refine data_sources.py and PyDataProvider2.py to make more readable
8 years ago
9f990d9059
Add unittest of the priorbox layer
8 years ago
454a1a2912
fixed a bug for demo/gan caused by batchNormLayer
8 years ago
1048aee0f7
Add input layer check
8 years ago
7b08a98a8f
do pre-commit
8 years ago
707a9c9bbd
Fix variable name and add the annotation
8 years ago
c0076084e2
Format the python file.
8 years ago
0eac399280
priorbox layer for ssd
8 years ago
79bcd05897
Merge branch 'develop' of https://github.com/baidu/Paddle into fix_data_sources
8 years ago
62b20ca033
refine data_sources.py and PyDataProvider2.py to make more readable
8 years ago
454ca01af3
Fix Travis-CI build error.
...
* error because merge of #711 . The old issue don't check pre-commit
hooks.
8 years ago
b993585925
Merge pull request #711 from Haichao-Zhang/input_types_check
...
adding input type check for python data provider
8 years ago
4afaaa4b92
Autoformat all files
8 years ago
b6aca330a2
Merge pull request #764 from emailweixu/multiple_parse
...
Correctly handling multiple calls to parse_config()
8 years ago
e9549cbb78
Change "Baidu, Inc" into "PaddlePaddle Authors"
8 years ago
3cf7337f30
Correctly handling multiple calls to parse_config()
...
To solve this, we maintain the list of DefaultNameFactory used in by trainer_config_helper,
and reset the state at the beginning of each parse_config call.
Change-Id: I13c7574dc8f0b6bc6f6b7c92eb425e2c52c926e8
8 years ago
4823075f95
Merge pull request #651 from Xreki/warpctc
...
Integrate warp-ctc as WarpCTCLayer, including unit test and layer interface.
8 years ago
16bebc4388
some tiny fix
8 years ago
78bdd32471
Merge branch 'develop' into warpctc
8 years ago
7c76d0eeac
config_parser new feature at ISSUE #726
8 years ago
8e9ac0cc55
adding input type check for data provider
8 years ago
7830893a83
fix conflict
8 years ago
1242e7944b
Merge branch 'develop' into fix-crf-weight-and-coeff-bug
8 years ago
25f1fbc491
Merge branch 'develop' into warpctc
8 years ago
d60d34ef3f
Fix bug in python/paddle/trainer/config_parser.py
...
The dims for crf and crf_decoding layers are wrong, i.e. height and width are swapped.
8 years ago
09a5b8bd4d
consolidate img_conv.conf in test_NetworkCompare
8 years ago
1d1a04c033
follow comments on rectangle CNN
8 years ago
4d487c6f35
Integrate warp-ctc as WarpCTCLayer, including unitest and layer interface.
8 years ago
7573205c68
follow comments on config_parser
8 years ago
96615fe329
merge develop, fix conflict
8 years ago
1a8fcc00de
fix the code style
8 years ago
86060772ed
Merge branch 'develop' of https://github.com/baidu/Paddle into config_parse_bug_fix
8 years ago
4ca9c3955e
minor changes
8 years ago
876fa1935d
Fix the config_parse.py if user does not set padding in the old config.
8 years ago
496d64ebdb
Support rectangle input for CNN
8 years ago
9115ab1c11
Merge pull request #450 from reyoung/feature/pre-commit-hooks-scripts
...
Feature/pre commit hooks scripts
8 years ago
bf6f690f31
Add ScalingProjection
...
out = w * input
where w is a parameter of size 1
Change-Id: Ife682d62323ceb1a20cbbf6269421b20a862d888
8 years ago
836d61382f
Update pre-commit-config
...
* Check all files by pre commit hooks
8 years ago
a1ba3f442f
format python code in python directory
8 years ago
ca0bb40c6e
Merge pull request #300 from QiJune/feature/sppnet
...
add SpatialPyramidPoolLayer c++ support
9 years ago
ebb153b067
Merge pull request #416 from yu239/hl_activetype
...
change the act.name for LinearActivation() to "linear" so that it won't fail in hl_activetype; also fix the hasinputsset in submodel
9 years ago
eb3bf9ea64
Merge pull request #360 from emailweixu/sum_cost
...
Add SumCost
9 years ago
36fa251756
'*' operator overload for LayerOutput
...
Making '*' support the multiplication between a scalar and LayerOutput
Also changing '+' to support adding between a vector and a scalar.
Change-Id: I7daf35590dc2b2f855a29d9ef43ac57979442e0f
9 years ago
9dd588b414
fix merge conflicts
9 years ago
8295eb91bf
Merge pull request #287 from gangliao/bilinear
...
Add bilinear interpolation layer
9 years ago
880774d141
change the act.name for LinearActivation() to "linear" so that it won't
...
fail in hl_activetype; also fix the hasinputsset in submodel
9 years ago
1c58e27fdc
fixed a bug in parse_conv in config_parser.py
9 years ago
af7a50c0d4
minor changes on deconv implementation and add protostr test for deconv layer
9 years ago
53e1629a43
Refactored imageSize in ConvBaseLayer to MathUtil
9 years ago
7a322df0a8
deconv implementation mionr changes in ConvBaseLayer.cpp and config_parser.py
9 years ago
d116b17f06
fix a small bug in ConvTransLayerBase in config_parser.py
9 years ago
fb20187aaa
deconv layer implementation modification following luotao1 comments
9 years ago
3d72e94939
rebase deconv implementation with develop branch and resolve conflicts with pull#218 commit 45c81a414f
9 years ago
5e4cc241ac
Revised deconv implementations according to luotao1
9 years ago
bda259bb18
added more test on convTrans layer and comments
9 years ago
70e44732c2
added convTrans test and python components
9 years ago
db569f293e
fix merge conflict
9 years ago
5ece5c96ad
add python wrap for sppLayer
9 years ago
ebad8e525d
Add SumCost
...
This allows user to implement any type of cost by summing over the output of non-cost layers.
Change-Id: Ic55aaabbf0c1299e70b8e48a0effcc91f8f5bd29
9 years ago
5a1e7dbc6a
Fix conflict
9 years ago
e802471c58
abstract outputSize function in CNN-related layers ( #314 )
9 years ago
bd38facada
Fix conflict
9 years ago
57348806b5
Follow comments
9 years ago
6c3a678c9a
Add elementwise math operations ( #343 )
...
* Add elementwise math operations
This allows use to use expressions like: y=log(1+exp(x))
Also added unittests for ActivationFunction
* Enforce keyword arguments for non-positional arguments
* Add LogActivation to doc
9 years ago
45c81a414f
Add job=time in trainer, refine cudnn_conv to reduce gpu memory and speed up training. ( #218 )
...
* Add benchmark for PaddlePaddle, tensorflow and caffe
* ConvProjection to reduce memory for goolenet
* Add unit test for ConvProjection.
1. unit test in test_LayerGrad.
2. compare the ConvPorjection and CudnnConvLayer, also compare the concat_layer+img_conv_layer and concat_layer_conv_projection.
* Reduce cudnn_conv memory and add benchmark document.
1. Use TmpMatrix as the workspace in cudnn_conv to reduce gpu memory. It reduce lots of memory.
2. Add benchmark document.
3. fix smallnet_mnist_cifar.py in paddle.
* Add job=time and refine cudnn_conv to reduce gpu memroy and speed up
* Refine cudnn_conv and shared biases operation in concat_layer and mixed_layer.
* follow comments
* follow comments
* Use unique_ptr to prevent memory leaks in CudnnConvLayer.
9 years ago
fd4eeaf59c
Merge conflict with maxout layer
9 years ago
ddfff3a7fd
Add bilinear interpolation layer
9 years ago
3dd8c9bea4
add maxout layer, including interface and unittest ( #229 )
...
* add maxout layer, including interface and unittest
* follow maxout comments
* auto setting channels
* fix unittest bug in test_RecurrentGradientMachine
9 years ago
2f82d72ede
Fix bug in yield dictionary in DataProvider. ( #197 )
...
* Fix bug in yield dictionary in DataProvider.
* Also make virtualenv work in Paddle.
9 years ago
6d21ecef54
add cost_type constraint to weighted_cost interface ( #206 )
9 years ago
191fafe355
support rectangle padding, stride, window and input for PoolProjection ( #115 )
...
* support rectangle padding, stride, window and input for PoolProjection
* Follow comments.
1. Remove start
2. refine img_pool_a/b.conf for test_NetworkCompare
3. Split unit test
* Modify the test in img_layers.py
9 years ago
d130d18146
Complete unittest for trainer_config_helpers. ( #108 )
...
* Fix lots of trainer_config_helpers bug, and complete unittest for `layers.py`
9 years ago
8e957df4b2
fix bug in dotmul_operator's api and anotation ( #99 )
...
* fix bug in dotmul_operator's api and anotation
* update rnn document
* remove redundant info of projection and operator in layers.py
9 years ago
b130ba7328
fixed bugs in conv_operator; add calc_output_size ( #101 )
9 years ago
2c5a6ac095
Optional fields to shrink generated proto size ( #93 )
...
* remove unnecessary field set in ParameterConfig, Evaluators, etc
9 years ago
159dd8331c
split dotmul_projection and dotmul_operator ( #87 )
...
* split dotmul_projection and dotmul_operator
* bug fix in outputsize checking for mixed layer
9 years ago
90b9cba7fe
Add min_pool_size, Add default value of should_shuffle ( #70 )
...
* min_pool_size would be infinite by default.
* add unittest for min_pool_size
* Fix bug in can_over_batch_size
* add unittest for can_over_batch_size
* Add DEFINE_PROVIDER_EX
* Add default value of should_shuffle
* When training, the default value of should_shuffle is True.
* When testing, the default value of should_shuffle is False.
* User a set a provider should_shuffle or not by pass it to `@provider`
* should_shuffle can handle a list of value, not just boolean
* Add input order mapping by using name
* Add unittest
* Add check to check input format.
* Default is close for speed reason.
* User could stop train when check error, or continue train without
this train sample.
* use deque instead of vector in generators pool, make erase
generator faster.
* Add chinese/english documentation
* Make should shuffle = false in unittest
* Add python files to depends.
9 years ago
05a97ab50e
Further fix the memory for Hierarchical RNN
...
Sequences should be sorted according to the number of subsequences they have.
9 years ago
c7762da330
Merge branch 'master' into merge_icode
9 years ago
699d5f2638
modify RecurrentGradientMachine to support unequal length inputs
...
* modify RecurrentGradientMachine to support hasSubSeq sequence inlinks with the same number of sentence but different number of tokens for each sentence
Change-Id: Ic71f00a4bb346b4fa93e650dfb4b1a0d8d2338b0
9 years ago
7481429f99
Change cos_sim to use CosSimLayer layer when size=1 and rename convex_comb_layer to linear_comb_layer
...
Also add an unittest for layers.py, currently for a few things. More need to be added later.
9 years ago
7dbc092c87
fix cudnn version number for batch norm. ( #71 )
...
* fix CUDNN_VERSION for backward of CudnnBatchNormLayer
* fix cudnn version number for BatchNorm
9 years ago
6a873f505f
fix the layer name error when stacking RecurrentGroups
...
* the last layer in the stack already has all the suffixes
9 years ago
b72beee4f0
fix dash and space bug,
...
ISSUE=4586495
git-svn-id: https://svn.baidu.com/idl/trunk/paddle@1408 1ad973e4-5ce8-4261-8a94-b56d1f490c56
9 years ago
Yancey
dangqingqing
Luo Tao
wangyang59
Yuan Gao
gaoyuan
Yu Yang
Peng Li
LCY-Seso
Helin Wang
Liang Zhao
xuwei06
wangkuiyi
Yi Wang
王益
Haonan
zhangruiqing01
caoying03
qiaolongfei
hedaoyuan
wangyanfei01
Yi Wang
Yiqun Liu
hanchao
zhanghaichao
qijun
Haonan
gangliao
emailweixu
Haichao-Zhang
zhangjinchao01