b3463bfffc
Merge pull request #3851 from luotao1/no_kernel_op
...
package a new USE_NO_KERNEL_OP for USE_OP_ITSELF
8 years ago
86655cb963
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_add_axis
8 years ago
e76fa85cc8
WIP
8 years ago
9a3c69c268
Revert "Remove `grad_op_type` in `REGISTER_OP`"
8 years ago
32d7e61880
Fix some bugs.
8 years ago
447033296d
Make some operator correctly handle gradients for multi inputs.
8 years ago
1a615b486f
Gpu test of im2col.
8 years ago
740c8ba12a
remove scatter_op.cu/gather_op.cu as they support only_cpu now
8 years ago
3f555001db
Add im2col test.
8 years ago
5b93ac7778
package a new USE_NO_KERNEL_OP for USE_OP_ITSELF
8 years ago
a60128aeb2
move identity_op from scale_op.cc to be a single file
8 years ago
58369d5c48
Merge branch 'develop' into op_refine
8 years ago
409ac4a390
Merge pull request #3819 from jacquesqiao/add-getop-to-ctx
...
add op() to InferShapeContext
8 years ago
f807807ed1
Fix the namespace of the im2col functor.
8 years ago
45c8f9b282
Add context parameter and math namespace.
8 years ago
abfac74c75
Merge branch 'develop' of https://github.com/baidu/Paddle into conv_op
8 years ago
d323831a0b
add GetAttr to InferShapeContext
8 years ago
79b1f33a02
Remove `grad_op_type` in REGISTER_OP
8 years ago
848c317aa8
update gpu code
8 years ago
6fcdc91638
add op() to InferShapeContext
8 years ago
c657537b4f
Correct some typos.
8 years ago
b1a185524f
Fixed SEGFAULT of dropout operator in GPU.
8 years ago
9a44f3d6da
Add dropout operator.
8 years ago
91215bcef9
Fix a bug causing wrong gradient results in cos_sim op.
8 years ago
3970f240b3
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix-3736
8 years ago
c33ddc74c1
Fix some bugs, add more unittests.
8 years ago
9d2909be2a
rename add_op to add_two_op
8 years ago
ed72af48ce
Add cos_sim op.
8 years ago
7f27b5e1b6
refine paddle/operator/CMakeLists.txt
8 years ago
7736992187
Simplify the CMakeLists for paddle/operator and paddle/pybind
8 years ago
6bef079660
Follow coding style and move reshaping operation to paddle tensor.
8 years ago
f8b885f27f
Using EigenTensor to reshape tensor.
8 years ago
a4df3f5bd8
Finish framework of squared_l2_distance_op.
8 years ago
2db7dedea9
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into pad_op
...
Conflicts:
paddle/pybind/pybind.cc
8 years ago
26cec83901
Add pad op
8 years ago
2d707e32c8
Refine the comments.
8 years ago
e967645c15
Refine the gpu code.
8 years ago
f7be9cb97a
Refine the cpu code.
8 years ago
6efbe2ff43
Merge im2col functor.
8 years ago
4f0c071e49
refine backward
8 years ago
bfeecfd3d2
fix problems
8 years ago
6f235553fd
scatter op fixed
8 years ago
97649bf9b2
fix codes in scatter
8 years ago
c5e28dd1a0
scatter check in
8 years ago
3663bd881d
Merge pull request #3620 from qingqing01/lookup_table
...
Add a lookup table op and a CUDA helper.
8 years ago
12864f1420
register rowwise add gpu kernel
8 years ago
aafeff0fde
Merge branch 'develop' into lookup_table
8 years ago
11f4283f09
Merge pull request #3634 from reyoung/feature/minus_op
...
Feature/minus op
8 years ago
72d29186bb
reduce some compile warning
8 years ago
fe480b9ebe
Merge branch 'develop' into lookup_table
8 years ago
25e6c014a4
Merge branch 'feature/identity_op' into feature/minus_op
8 years ago
b53f4e0be7
Merge branch 'develop' of github.com:baidu/Paddle into feature/identity_op
8 years ago
3a6f20025a
Merge branch 'develop' into feature/extract_op_info_into_op_info.cc
8 years ago
d8ea560eb2
resolve conflicts
8 years ago
f188e22b33
Remove set functor and add comapre_grad test
8 years ago
0a0f194836
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
a8d072c769
fix bug.
8 years ago
31f59d2b31
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into lookup_table
8 years ago
5810d63ffc
Merge pull request #3617 from QiJune/fix_code_style
...
fix code style
8 years ago
f931140663
Merge pull request #3501 from qingqing01/cross_entropy
...
Implement GPU kernel for cross entropy operator.
8 years ago
1795e57671
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into lookup_table
8 years ago
0f3b9e4112
lookup table op, cuda helper and set functor
...
1. finish lookup table CPU and GPU kernel
2. Add some cuda helper
3. Add some math funtor
8 years ago
5d0d44a018
use clang-format
8 years ago
9b4a9da9f2
fix code style
8 years ago
ce723af062
Merge pull request #3585 from reyoung/feature/rename_to_append_op
...
Rename `Net::AddOp` to `Net::AppendOp`
8 years ago
950cc60d2b
Add minus
8 years ago
7443b2e481
Merge pull request #3596 from QiJune/implement_random_function
...
refine random related operators
8 years ago
dc5f0dbc32
remove opregistry in gather function
8 years ago
aff90d8ee7
fix gpu build error
8 years ago
1918ad8759
fix gpu build error
8 years ago
36e8e72566
expose random seed to users
8 years ago
53e71b44f4
gather op bp passed
8 years ago
f3df105455
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
d3c29e9d16
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
a75a638fb1
format Copyright
8 years ago
950dbde56c
fix rowwise add grad op
8 years ago
d3f219aa99
Change IdentityOp to ScaleOp
8 years ago
b054392e2a
fix gaussion op bug
8 years ago
08c987d7c0
use dynload curand
8 years ago
2f47f35b3e
fix gpu build error
8 years ago
7c274dc0a1
use curand
8 years ago
c108d6108c
Identity operator and its gradient
8 years ago
d525abed95
refine random related ops
8 years ago
59b3df31aa
Extract OpInfo into a library
...
Fix cycle dependencies, Fix #3583 .
8 years ago
f1e5533541
Rename `Net::AddOp` to `Net::AppendOp`
...
Fix #3582
8 years ago
a8863a8d5b
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into cross_entropy
8 years ago
8f6c8780a5
Replace functor by function.
8 years ago
78553768ba
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
b59002daef
"fix math gemm lda order error"
8 years ago
514398c0b1
"delete unused comment"
8 years ago
0cf5bdec56
"tensor mutable data"
8 years ago
2799da6634
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into mul_op
8 years ago
ab270c3899
Merge pull request #3183 from dzhwinter/add_op_gradient
...
"add rowwise add backward op"
8 years ago
82b820e97b
fix rowwise_add_grad_op
8 years ago
70285cce32
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into cross_entropy
8 years ago
230e613c6f
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix-sgd
8 years ago
8b3d33a055
fix-sgd
8 years ago
cef27dab47
"add fixl"
8 years ago
50cf127eea
"change Output to Input"
8 years ago
c332e4ee25
"relauch the ci"
8 years ago
448d4db5c1
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_op_gradient
8 years ago
597ac21504
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into mul_op
8 years ago
7b4b9d3e09
"format style"
8 years ago
940b4a9c82
Merge pull request #3542 from Canpio/change_name_of_AsNoGradient
...
Rename `AsNoGradient` of VariableBuilder to `NotInGradient`
8 years ago
88360100bf
Merge remote-tracking branch 'origin/develop' into mul_op
8 years ago
62aedcee31
Merge pull request #3553 from reyoung/feature/unittest_for_mean_grad
...
Add MeanOp's Gradient Test And Fix Mean Op Gradient
8 years ago
7f8c3f8214
Add MeanOp's Gradient Test And Fix Mean Op Gradient
8 years ago
c68bfc30a8
Merge pull request #3476 from qingqing01/bp_test
...
Compare the gradient consistency between GPU and CPU calculations
8 years ago
1808f888b9
Merge pull request #3538 from reyoung/feature/remove_shared_ptr
...
Feature/remove shared ptr
8 years ago
812a64c083
Merge pull request #3534 from gangliao/mem_release
...
FIX: Release CPU/GPU memory at the end of the Program
8 years ago
1365f2d15e
Remove R-Value reference in AddOp
...
Fit Google C++ Style
8 years ago
5181aefc6b
tune max relative error for sigmoid op unit test.
8 years ago
18dcc1c0bc
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into bp_test
8 years ago
225579b9d9
Remove own for add_op
...
* add_op could take a unique_ptr or a const reference. If unique_ptr is
taken, the NetOp will take care of that operator's life cycle. If a
const reference is taken, that op will be Cloned.
8 years ago
d8560ec2e8
Fix scatter_test
8 years ago
ac02fb82d7
FIX: tensor memory must be gaven back to buddy allocator for free
8 years ago
3484874278
Rename `AsNoGradient` of VariableBuilder to `NotInGradient`
8 years ago
ee29d1b6e2
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
6075928d55
gather op added
8 years ago
f15e083098
Remove std::shared_ptr in Python & C++
...
* Also simplify pybind implementation by using OperatorBase as holder
type.
8 years ago
1543eeb4ce
init
8 years ago
8c653ba76a
Complete remove std::shared_ptr
8 years ago
c7f2532594
Merge branch 'develop' of github.com:baidu/Paddle into feature/clone_op
8 years ago
1425387570
Using unique_ptr instead of raw ptr
...
Fit google C++ style
8 years ago
29d892c13c
Add Clone Method For OperatorBase
...
* Clone method will create a new object instance, which is as same as
itself.
* This is the first step to remove shared_ptr for OperatorBase
8 years ago
4d2adab772
gather op added with python unittest
8 years ago
323d4233f3
gather op added with python unittest
8 years ago
26475cd9ba
Use clipping log in cuda kernel, making it same with CPU.
8 years ago
9eaef75397
RNN backward create ( #3490 )
...
* insert rnn's backward into Backward()
* add device_context into backward_test
8 years ago
49aa2c042c
Implement GPU kernel for cross entropy operator.
8 years ago
e256bfaf28
"update paddle enforce"
8 years ago
4ab36a71c4
"fix error"
8 years ago
4c9699c5f8
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into mul_op
8 years ago
f6bffd4e1f
gather_op modified
8 years ago
2a42a73db1
modify gather_op with test
8 years ago
caaa5f86b9
gather op added
8 years ago
af1eb31afc
add as an operator
8 years ago
33d502e73b
Merge pull request #3461 from tensor-tang/mklso
...
add shared mkl lib in whl
8 years ago
0079fa3256
Rnn make stepnet member ( #3469 )
...
* make stepnet member
* add pybind support
* fix Inputs Outputs
* remove unique_ptr
8 years ago
e9eee6f785
"polish words"
8 years ago
d9400243d9
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into public_to_protected
8 years ago
973618b6ab
Merge pull request #3371 from zchen0211/develop
...
scatter update implemented
8 years ago
5d142b0c78
Fix compile errors
8 years ago
864b00cdf0
Replace `EmptyOp` with `NOP`
8 years ago
3e6e5c9286
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refactor_registry_macro
8 years ago
7aaf019a14
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
2be3d32711
use shared lib when mkl
8 years ago
84d6434d53
Compare the gradient consistency between GPU and CPU calculations.
8 years ago
12ee501485
"fix operator grad config"
8 years ago
0c96c99746
change pybind and net_op_test
8 years ago
2f74e60865
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into public_to_protected
8 years ago
43ba24e051
Merge remote-tracking branch 'origin/develop' into add_op_gradient
8 years ago
e0395a53e9
"remove unused commented code"
8 years ago
632b320e9d
"refine argument with new style "
8 years ago
426d7328a0
Merge remote-tracking branch 'origin/develop' into mul_op
8 years ago
ffbb0be21f
Merge pull request #3444 from reyoung/use_ctor_create_op
...
Using constructor to create an operator.
8 years ago
b2e3824e41
change operator
8 years ago
c41862d2f5
merge baidu/develop
8 years ago
fee8aad954
Merge remote-tracking branch 'baidu/develop' into port_blas
8 years ago
daaa45b427
Merge branch 'develop' into use_ctor_create_op
8 years ago
960a525550
fix gpu build error
8 years ago
2ec8dab4c7
follow comments
8 years ago
f09cb657e6
Follow comments from WangYi
8 years ago
4a604c2651
Polish Our code by YuYang's review
8 years ago
b63e1c6d8a
"op name"
8 years ago
7a31d72d6b
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refactor_registry_macro
8 years ago
11c35605fc
Remove empty constructor for operator
8 years ago
0b1052fca5
Get `DEFINE_OPERATOR_CTOR` Back to code
8 years ago
d6d46412ef
Merge branch 'develop' of github.com:baidu/Paddle into final_fixes
8 years ago
2ea2fbea1b
Merge REGISTER_OP and REGISTER_GRADIENT_OP
8 years ago
8e36ade03c
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
38f4b1d59e
Merge pull request #3430 from wangkuiyi/add_operatorbase_constructors
...
Add constructors to OperatorBase and all sub-classes
8 years ago
3e11e4c6b9
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refactor_registry_macro_dev
8 years ago
65bd7c77e4
Update
8 years ago
f784741d4a
Refine macro
8 years ago
bfde09b8e0
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into develop
8 years ago
9430bc3207
fix all bugs
8 years ago
580445a430
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refactor_registry_macro
8 years ago
96fc9e7d1b
Merge pull request #11 from reyoung/fix_python_tests
...
Fix python unit tests
8 years ago
9a592ec3aa
Remove FC Op, since it should be added in Python side
8 years ago
c99f84aced
Fix python unit tests
8 years ago
37aa4b98ff
refine unittest
8 years ago
c2631ebf6f
add unittest
8 years ago
03d0040c59
gather warning fixed
8 years ago
554e08ac7a
fix build error: 'type name' declared as function returning a function
8 years ago
95df4700d4
Merge remote-tracking branch 'baidu/develop' into port_blas
8 years ago
7fab7ddd7d
Merge branch 'develop' of github.com:baidu/Paddle into feature/refactorize_framework_proto
8 years ago
227e09b208
Merge branch 'develop' of github.com:baidu/Paddle into feature/fix_random_op_link_error
8 years ago
688c43b104
format code
8 years ago
7202f425dd
Merge branch 'refactorize_framework_proto' into feature/refactorize_framework_proto
8 years ago
4591110201
Fix gaussian_random_op compile error
...
* Should always use `dynload::` for cuda function.
* Fix cublas.h without DSO load.
8 years ago
5f1081d83d
fix bug in dynload
8 years ago
c326aae0cf
Fix code style in gaussian_random_op.cu
8 years ago
c5a7471e93
add math_function_test
8 years ago
331b304a18
merge baidu/develop
8 years ago
52b52ba80c
fix gpu build error
8 years ago
de967fcefe
set gemm support continuous memory now
8 years ago
7eb07b3399
fix conflict
8 years ago
8de4e3bdd6
disable gpu implementation temporarily
8 years ago
2ddb11222a
"on hold"
8 years ago
e81a5f870f
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into refactor_registry_macro
8 years ago
bc1459cefb
Unable macro `USE_OP_GRADIENT` temporarily
8 years ago
39f14f1dd6
scatter update implemented
8 years ago
57c0978411
Fix bug
8 years ago
7c8e5c3b63
Merge pull request #3191 from zchen0211/develop
...
gather function only
8 years ago
d4e4cebf5f
fix all coding-style problems
8 years ago
56faf51343
Merge pull request #3060 from dzhwinter/random_op
...
Random op
8 years ago
7307b439e1
fix gpu build error
8 years ago
030f4302ed
Merge branch 'develop' of github.com:baidu/Paddle into feature/refactorize_framework_proto
8 years ago
840d0c7402
Remove unnecessary C++ operator test
...
They are tested in Python
8 years ago
78af6e6011
Add OutputVars method to get all outputs or outputs without intermediate
8 years ago
5a59111700
Modify rnn op unit test after refactoring framework proto.
8 years ago
b228b463fa
Make const variables in operator.h fit google style
...
* No POD instance is forbidden in global scope.
See
https://google.github.io/styleguide/cppguide.html#Static_and_Global_Variables
8 years ago
6bac3e17b5
"remove unused test net modified"
8 years ago
a821fec158
merge baidu/develop
8 years ago
200e3e2c6b
"change namespace prefix"
8 years ago
a0e195a175
Merge remote-tracking branch 'origin/develop' into add_op_gradient
8 years ago
6159f5db14
code style fix
8 years ago
9544068543
Resovle conflicts manually
8 years ago
689d5ee157
solve conficts
8 years ago
6fc6647c31
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
b7ee1e7d9c
"backward check todo"
8 years ago
789d6ed9b7
Merge remote-tracking branch 'origin/develop' into add_op_gradient
8 years ago
22f03c3981
"fix clang format"
8 years ago
6535a7b001
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
91e8aadaef
Merge branch 'develop' into type_alias
8 years ago
e31a469ee0
add gradient test framework ( #3226 )
...
* init grad op checker
* can run
* add GradeChecker class
* use get_numeric_gradient
* refine code
* add softmax and cross entropy auto grad test
* use close to judge op_grad and numeric_grad
* add cpu and gpu compare
* add comments
* add support_gpu
* fix allclose
* fix name error and symplify code
* optimize gradient checker
* add test_cross_entropy_op
* update gradient_checker.py
* optimize code
* use random.uniform instead of random.random
* fix type bug
* optimize check_grad
* put SupportGPU into OperatorBase
* typo
8 years ago
23ac845921
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
2c553e4fb6
"fix clang format error"
8 years ago
dba618c036
Make Compile Pass
...
* Although backward_test/rnn_test is not pass, just comment them.
8 years ago
cf92472822
clang format
8 years ago
555af4d0f6
"format code"
8 years ago
8804b242d8
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
7945572cee
Update fill_zeros_like_op.cu
8 years ago
e2c08d286f
"keep style same with uniform operators"
8 years ago
7c2508109e
Follow comments, change auto -> auto*
8 years ago
174a3b3ec6
Merge branch 'develop' of github.com:baidu/Paddle into feature/uniform_random_op
8 years ago
b32fcf2854
merge origin/develop
8 years ago
7e830116a7
Try make pass
8 years ago
2af35002f7
fix some enforce ( #3301 )
...
* fix some enforce
* remove compatible_type to avoid compile error
* remove shared_ptr
* fix tensor error msg
8 years ago
b861c01963
Merge branch 'develop' into feature/uniform_random_op
8 years ago
662aeed28f
Update operator/CMakeLists.txt
8 years ago
7e37471318
modify
8 years ago
2847667620
fix bug
8 years ago
5d2b3fb09d
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_add_FillZerosLikeOp_test
8 years ago
f50730fda1
"fix recurrent op test"
8 years ago
e5847924f3
Merge branch 'develop' of github.com:baidu/Paddle into feature/uniform_random_op
8 years ago
79e76ea1ed
"fix cuda error"
8 years ago
4bbd05fd72
check INFINITY in cross_entropy ( #3287 )
...
* check INFINITY in cross_entropy
* fix error
* use onehot_cross_entropy without GPU kernel
* add support_gpu
* fix allclose
* fix name error and symplify code
8 years ago
1752b6a1b6
Merge remote-tracking branch 'baidu/develop' into port_blas
8 years ago
376737d8b8
Follow comments
8 years ago
876e9d2bc2
"sigmoid confict resolved"
8 years ago
691615b146
Merge remote-tracking branch 'origin/develop' into type_alias
8 years ago
fcd6f64b98
"redefine random op"
8 years ago
58561d8f56
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
0c37705ddc
Use thrust to implement uniform_random
8 years ago
bd369c3575
"remove type alias header file"
8 years ago
72fb86a284
"remove type alias done."
8 years ago
9e25988f73
"net op alias"
8 years ago
610801b57c
"remove a lot alias"
8 years ago
6b23b91cea
"remove alias to more operators"
8 years ago
efa50a4525
Merge pull request #3181 from qingqing01/sigmoid_bp
...
Sigmoid backward implementation.
8 years ago
bee9596759
"remove type alias"
8 years ago
6b12c697ff
handle mkl
8 years ago
7665bdba13
Rnn forward logic test ( #3291 )
...
* finish forward debug
8 years ago
0815935916
fix typo error
8 years ago
5703eb50fa
add .clang-format file
8 years ago
01a198a584
merge baidu/develop
8 years ago
97d8175a5e
add global matmul function for Tensor
8 years ago
9f490c7748
update to develop branch.
8 years ago
e376bda42c
Add uniform random operator
...
It can be run both CPU/GPU. configure attributes are:
* min: the min value of uniform random
* max: the max value of uniform random
* dims: the dimension of output tensor
* seed: the random seed of uniform random. 0 means generate a seed each
time.
8 years ago
0f8c9dbe61
device context pointer
8 years ago
2447c34a0e
merge origin/develop
8 years ago
c72253127f
crossentropy grad op ( #3186 )
...
* init cross entropy graident
* add crossentropy grad op
* remove details
* fix static compile
8 years ago
6b3e0b786d
gather function with test passed
8 years ago
b943874f16
move namespace rnn to a directory ( #3261 )
...
* move namespace rnn to a directory
8 years ago
9620df4464
Reformat paddle/operators/* strictly following Google Style Guide
8 years ago
301a21d8fe
cpplint recurrent_op*
8 years ago
a07cef338c
Merge branch 'develop' of https://github.com/paddlepaddle/paddle into cpplint_errors
8 years ago
a40b755b6a
Add explicit to some constructors
8 years ago
9b0fce516a
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_add_FillZerosLikeOp_test
8 years ago
9d68c252e9
test
8 years ago
eae5c94944
Merge branch 'develop' of https://github.com/paddlepaddle/paddle into const
8 years ago
bf740a3fe7
merge baidu/develop
8 years ago
5d7e8bfb5c
fix bug
8 years ago
c6186120c3
fix softmax_op code line > 80
8 years ago
430b421906
Fix conflicts
8 years ago
b58725bd51
Add cpplint for *.h and cuda *.cu
8 years ago
ff19223e8d
Reforamt
8 years ago
ddb29b6ca6
Move constants from framework::OperatorBase to framework::
8 years ago
8ff3590eda
fix op name
8 years ago
df84e39a00
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_add_FillZerosLikeOp_test
8 years ago
0bd49a50ce
move net_op to operators/ ( #3201 )
...
* move net_op to operators
8 years ago
d953611e86
Softmax grad op ( #3164 )
...
* init softmax grad op
* add compute code
* export Backward to python
* update test ,export op.type to python
* update python test, fix compute bug
* update unit test
* use eigen
* optimize eigen code
* add gpu test
* register softmax_grad GPU kernel and fix test bug
* typo
* follow comments
8 years ago
f190a79538
fix gpu build error
8 years ago
22dac40c3a
add gemm for both cpu and gpu
8 years ago
4eb25b4793
update code.
8 years ago
eef55ca700
remodify
8 years ago
fa66cc77ff
Mean Op Merge Develop
8 years ago
2b35fca18f
gather modify
8 years ago
3064ae3304
Merge pull request #3180 from reyoung/feature/MeanOpGrad
...
Add Gradient Operator for mean
8 years ago
017182c6b1
Merge pull request #3124 from qingqing01/rnn_infershape
...
Refine InferShape for recurrent_network_op
8 years ago
43528c4c08
fix conflicts
8 years ago
0802197924
gather and scatter-update added
8 years ago
cc6c33b8b2
export Backward to python ( #3174 )
...
* export Backward to python
8 years ago
691a00e3d2
Merge pull request #3178 from reyoung/feature/GetEigenDeviceReturnReference
...
Return Reference Instead Pointer to GetEigenDevice
8 years ago
264b644718
"add rowwise add backward op"
8 years ago
54f0d2608a
fix input size.
8 years ago
bd01cea1f3
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into sigmoid_bp
8 years ago
0560733c2e
Add sigmoid backward implenmention.
8 years ago
18d924173f
Add Gradient Operator for mean
8 years ago
b36205e296
Refine compute code in operators
8 years ago
7c49a4f31c
resovle conflicts.
8 years ago
02655a229e
Return Reference Instead Pointer to GetEigenDevice
8 years ago
341d1880ee
Merge remote-tracking branch 'baidu/develop' into op_gpu_test
8 years ago
81cc7a3227
merge baidu/develop
8 years ago
99b835cc15
resolve conficts.
8 years ago
3a8017d148
Merge branch 'backward_dev' into dev_add_FillZerosLikeOp_test
8 years ago
8bd73159c7
Add unittest for `FillZerosLikeOp`
8 years ago
6e0661cf9c
Change ContextKernel to ExecutionKernel
8 years ago
426f099932
Remove duplicate fc_op
8 years ago
c36a093c5e
Fix conflict
8 years ago
18aa0dcb2b
resolve conflicts.
8 years ago
9cc990741e
Merge branch 'develop' of github.com:baidu/Paddle into feature/backward
8 years ago
4f1061f799
Merge pull request #3133 from gangliao/tensor_scalar
...
ENH: Add EigenScalar
8 years ago
b94584cf4b
Rename recurrent_network_op recurrent_op.
8 years ago
79e89ef43b
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into rnn_infershape
8 years ago
8925295a4b
follow comments.
8 years ago
1e676f684d
Add mean op unit test in python
8 years ago
47690d6a7b
ENH: Add EigenScalar
8 years ago
e2fd2bd0d1
Follow comments and merge develop
8 years ago
65880f7e9c
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into scope
8 years ago
bc146e8f6d
Merge branch 'develop' of github.com:baidu/Paddle into feature/backward
8 years ago
5d134a03f1
Refine remove std::shared_ptr in Scope
...
* Make interface of Operator to `const Scope&`
8 years ago
61ebacbcd3
use operator context and infer context ( #3024 )
...
* use operator context
* optimize code
* update net infershape
* update InferShape
* disable override InferShape(scope) in OperatorBase
* change InferShapeImpl to InferShape
* add template to OperatorContext Input/Output
* merge Input InputVar, Output OutputVar
* change Inputs to MultiInput
* fix conflict
* fix MultiInput bugs and add unit test
* rename KernelContext to ExecutionContext
* clean code
* change InferShape to protected
* fix template bug
* refine code
* use InputVar instead of Input<Variable>
* typo
* optimize code
8 years ago
bb7f34594c
Add mean cpu/gpu op
8 years ago
6cfb9a3262
Refine InferShape for recurrent_network_op.
...
* the tensor only contains shape and does not hold memory when inferring shape.
8 years ago
61f94f0002
add EIGEN_USE_GPU macro to op.cu file
8 years ago
ae3a9c48ed
Add throw_on_error specialization for T*; Update all usages of Scope
8 years ago
933e55eb06
fix conflict
8 years ago
49739265c7
"fix register error"
8 years ago
47556689d4
"remove unused code"
8 years ago
572133400d
"update the compute kernel"
8 years ago
5c437c43c7
fix paddle/develop build error concerning rnnop
8 years ago
aee0d3ec5f
RecurrentOp implementation ( #2890 )
...
* add rnn op interfaces
* add Run
* rename state -> memory
* change state -> memory
* make compilable
* add .cc
* init test
* add op fake implementation
* add CreateStepNet and CreateScopes implementation.
* add TODO list
* init memory attributes.
* add LinkMemories
* add PlainNet fake implementation
* Use std::shared_ptr<Scope> in the OpRunContext.
* add test
* disable mutable_data
* finist segmentInput function
* enable mutable_data with a trick
* RNNOp test.
* enable LinkMemories with mutable_data
* update SegmentInput function with comments
* finish ConcatOutput function
* reformat inputs and attributes
boot_memories
* Refine unit test.
* Refine unit test.
* modify inlinks.
* add OpDesc to Net
* fix bug and update unit test.
* move step scopes from inputs to outputs
* fix merge conflict, update SegmentInput function
* add RecurrentOpProtoAndCheckerMaker.
* clean the codes
* Abstract GetStepScopes and GetMaxSeqLen function
* refine LinkMemories
* Refine code and add some comments.
* add backward core
* update for develop branch.
* add forward core
* add forward algorithm
* Add RecurrentGradientAlgorithm implenmention.
* use CopyFrom and Slice function in RecurrentOp
* add unit test for LinkMemories.
* fix unit test.
* use the latest tensor.h, solve conflict
* add maker
* move SegmentInput and ConcatOutput to details nameplace
* unit test for RecurrentGradientAlgorithm.
* apply OperatorBase
* apply net operator.
* move memorys to attributes
* add RecurrentGradientOp
* open test unit test in recurrent_network_op_test.
* revert some files.
* add RecurrentArgument and Link struct to simplify member variable.
* rename.
* move recurrent_op from framework to operators
* add RecurrentGradientOp Init
* fix name
* fix Link.interal/external name
* use namespace operators instead of framework
* clean the code
* use the latest add_op and mul_op, don't test backward now
* Remove ScopePtr and OperatorPtr
* add get_net to pybind
* add test_recurrent_op.py
* add random into gen_tensor
* update to develop branch and refine some code.
* add some comments.
8 years ago
01626be9b3
Merge branch 'develop' of github.com:PaddlePaddle/Paddle into rnn_varilen_design
8 years ago
226bf1dded
update more details
8 years ago
d3213e4ca3
fix pr
8 years ago
a74e798103
add an appendix to explain sequence info data structure
8 years ago
70bd07a0e1
Fix compile errors of FillZerosLikeOp
8 years ago
e0463acfc3
Rename PlainNet --> NetOp
8 years ago
a2dc9614ed
Add fill_zeros_like op
8 years ago
00615ebca2
Refine OpRegistry::AddInput/AddOutput
...
Remove bool argument, use a class to handle that.
8 years ago
b1b4364580
Rename PlainNet --> NetOp
8 years ago
69b1b26511
"cpu only macro"
8 years ago
11f9f5fb17
"fix const dependency hell"
8 years ago
984225ecf1
"fix operator"
8 years ago
2b3e362136
"template specialization link include"
8 years ago
efc119b43b
Add type_alias to import framework into ops
...
Make implement an operator less noisy.
8 years ago
30a47fe8a3
"link pybind11"
8 years ago
ddafe5ceeb
Merge branch 'develop' of github.com:PaddlePaddle/Paddle into rnn_varilen_design
8 years ago
32c15a2919
"random op test"
8 years ago
1328060abc
update2
8 years ago
d263ccefd6
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
6f80b5f1df
"move to template function"
8 years ago
a0a2f1bf2c
update
8 years ago
e8a0e92b5f
Fix compile error
8 years ago
197c82f30f
Rewrite grad_op registry macro of mul, sigmoid and softmax
8 years ago
77af58f8f7
Change gradient Op registry mechanism
...
OLD: op_type -> grad_op_creator
NEW: grad_op_type -> grad_op_creator
op_type -> grad_op_type
8 years ago
0d554f1dea
"add template fill function"
8 years ago
5f3bc2a44a
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into backward
8 years ago
c110f56574
Merge remote-tracking branch 'origin/develop' into random_op
8 years ago
5ad9474bf7
add random op
8 years ago
45072ed2cd
add more details
8 years ago
6129ab42f9
Merge pull request #2964 from Canpio/dev_refactor_tensor
...
Simplify Tensor implementation
8 years ago
4c33d55a32
add rnn_design.md
8 years ago
06acd6d0cb
add unittest for some basic OpKernels
8 years ago
06358129a2
Fix compile error
8 years ago
ba79050923
Merge remote-tracking branch 'baidu/develop' into implement_basic_OpKernel
8 years ago
6cd94cc73f
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into dev_refactor_tensor
8 years ago
2685765905
add cross-entropy-op ( #2965 )
...
* add cross-entropy-op
* add infershape and compute
* implement Infershape and compute of onehotcrossentropy op
8 years ago
6f0539246c
Merge remote-tracking branch 'origin/develop' into backward2
8 years ago
81a352aff7
"test fc without gradient"
8 years ago
14424f314c
"use built-in operator"
8 years ago
f46961e223
Merge remote-tracking branch 'baidu/develop' into implement_basic_OpKernel
8 years ago
e4984f13e9
fix tensor usage in sgd-op
8 years ago
a6347daa05
fix clang compile warning
8 years ago
3014662d4a
Merge branch 'implement_basic_OpKernel' of https://github.com/QiJune/Paddle into implement_basic_OpKernel
8 years ago
e3b27d1998
Add sgd op ( #2950 )
...
* a simplest SGD op
8 years ago
736d078cbf
replace Tensor::tensor to EigenTensor::From
8 years ago
8ad9006dd5
Merge remote-tracking branch 'baidu/develop' into implement_basic_OpKernel
8 years ago
d81084939b
Merge pull request #2953 from wangkuiyi/tensor_type_to_eigen
...
Refactorize Tensor to Eigen convesion
8 years ago
55d301722f
Simplify Tensor implimentation
...
ATTENTION: some interfaces changed:
1. void Tensor::set_dims(const DDim& dims) ==> void Tensor::Resize(const DDim& dims).
2. void Tensor::ShareDataFrom(const Tensor& src) ==> void Tensor::ShareDataWith(const Tensor& src)
3. DDim Tensor::dims() const ==> const DDim& Tensor::dims() const
8 years ago
14cfb8c262
fix gpu build error
8 years ago
d9fa6159b7
add Flatten method to EigenVector
8 years ago
00ed564307
Update
8 years ago
b6c075527c
implement some basic OpKernel
8 years ago
d3373c5b85
Skeleton Of fully connected operator
...
* Fc operator is a grouped operator, which combined by may internal
operators.
* InferShape & Run a FC operator in Python.
8 years ago
d344f6702c
Merge branch 'develop' into tensor_to_EigenTensor
8 years ago
122e83e36c
Fix unittest
8 years ago
73a9f0f25d
Merge develop
8 years ago
ad8fa77cce
Merge branch 'develop' into feature/add_some_skeletons_of_ops
8 years ago
5847b96a61
Follow comments, correct implement `DDim::size()`
...
* Also fix unit test
8 years ago
87189665d9
merge baidu/develop
8 years ago
2a03e3808d
set correct place for output tensor
8 years ago
a0caf23430
Op varient inputs ( #2901 )
...
* add inputs
* add ut for multiple inputs
* fix AddToLayer
* op_desc -> op_proto
* CreateArgumentOffsetMap -> CreateInOutOffsetMap
* move CreateInOutOffsetMap from OperatorBase to op registry
* arg_idxs_ -> in_out_idxs_
8 years ago
1ed237c11e
Add skeletons of `mul`, `rowwise_add`, `sigmoid`, `softmax` ops
...
* Implement InferShape and register them, give a stub Kernel method
by LOG(INFO)
8 years ago
6dc567a52e
merge baidu/develop
8 years ago
38310f9349
Refine CMake dependencies graph
8 years ago
9e3a9eb233
merge baidu/develop
8 years ago
d649dbf442
implement add_op kernel
8 years ago
d3a749a5bf
CMake `op_library` function
...
* It is used to create an operator library. It handles to split CPU and
GPU sources and links operator common libraries.
* It also give a reasonable warning and error when operator developer
not correctly implement an operator.
* Warning for lack of GPU kernel.
* Same interface as `cc_library` to make code style consistent.
8 years ago
569f7e83a6
FIX: cppint code style
8 years ago
bac1426d47
add_op kernel implementation
8 years ago
0b0b3ba16a
Merge remote-tracking branch 'baidu/develop' into tensor_to_EigenTensor
8 years ago
e588730147
change op to operators
8 years ago
79b70c2d69
Follow comments
...
* Convert `op` --> `operators`
* Remove AddType in OpProtoMaker, because type is part of registry.
* Rename CPU_OR_GPU --> DEVICE_TYPE in registry macro.
8 years ago
71e2a94310
Merge remote-tracking branch 'baidu/develop' into tensor_to_EigenTensor
8 years ago
0ff8192072
Add OperatorWithKernel class
...
* User can register OpKernel to its Ops. The OpKernelMap saved in
OperatorWithKernel. Each Op which inherits OperatorWithKernel will
use `OpKernel::Compute` instead of Run.
8 years ago
37bcc0860a
merge baidu/develop
8 years ago
a2e5f652d3
add operator base ( #2725 )
...
Add OperatorBase.
issue: https://github.com/PaddlePaddle/Paddle/issues/2790
Paddle design the Operator with Kernel. OperatorBase has no type and device information when create, One operator can have multiple kernels, Operator will choose a kernel to run according to context. The kernel should be bind to Operator before or during Operator running.
8 years ago
958511160b
add simple add_op_functor
8 years ago
Tao Luo
fengjiayi
hedaoyuan
dangqingqing
Qiao Longfei
Xinghai Sun
yangyaming
wanghaoshuang
qijun
zchen0211
Yu Yang
dongzhihong
gangliao
superjom
Yan Chunwei
Yi Wang
tensor-tang
Yi Wang
dzhwinter
Yi Wang