36dce65bb3
Take DataType and VarType apart
...
test=develop
6 years ago
cc0ae1f1a1
refine with comments
...
test=develop
6 years ago
a275fd6e0c
Merge branch 'develop' into runtime_context
6 years ago
2579ade45f
Add cpu_quantize_pass for C-API quantization ( #16127 )
...
* Add cpu_quantize_pass for C-API quantization
test=develop
* add cpu_quantize_pass test
* fix lint: add include memory unorderd_map and unordered_set
test=develop
* fuse_relu 1
test=develop
* tuned 2 without squash
* fixes
test=develop
* remove unused vars
test=develop
* refactored
test=develop
* fix lint c-style cast -> C++ style cast
test=develop
* remove QuantMax and c style casts
test=develop
* last usage of QuantMax removed
test=develop
* Fix Analysis Predictor UT
Check if memory_optimize_pass has already been added
to the analysis config before adding a new one, so
that it is not added multiple times.
test=develop
* change map to unordered_map
fix the forgotten part of cpu_quantize_pass_tester.cc
test=develop
* removed quantized attribute
* fixed cpu_quantize_pass_tester and op attr comments
test=develop
* removed redundant line
test=debug
* removed gmock
test=develop
* fix after merge
6 years ago
86e912c544
Fix windows compiling ( #16230 )
...
test=develop
6 years ago
1b59bed989
Merge branch 'develop' into runtime_context
6 years ago
6ce25c99a0
Merge branch 'develop' into runtime_context
6 years ago
8ad672a287
Support sync batch norm. ( #16121 )
...
* Support Sync Batch Norm.
* Note, do not enable it in one device.
Usage:
build_strategy = fluid.BuildStrategy()
build_strategy.sync_batch_norm = True
binary = fluid.compiler.CompiledProgram(tp).with_data_parallel(
loss_name=loss_mean.name,
build_strategy=build_strategy)
6 years ago
ca392c7e97
Implement infer var type context
6 years ago
b9252f3df8
Add cpu_quantize_squash_pass for C-API quantization ( #16128 )
...
* Add cpu_quantize_squash_pass for C-API quantization
test=develop
* add cpu_quantize_squash_pass teste
* fix lint: add include memory unorderd_map and unordered_set
test=develop
* lint fix 2
* fixes
test=develop
* refactored
test=develop
* fix windows ci
test=develop
6 years ago
d94fd97230
add runtime_context_cache_pass
...
test=develop
6 years ago
41b8cf0bae
Merge pull request #16162 from wzzju/fix_nan_static_quant
...
Fix NaN bugs for static quantization strategy (mutil-cards train).
6 years ago
5685a48c23
Add some fixme. test=develop
6 years ago
ac6ef06ffa
Add the Clone method in Graph. test=develop
6 years ago
01eddf125c
Not add graph copy construction method. test=develop
6 years ago
1b9c8d5f06
add clone function for IrGraph. test=develop
6 years ago
40f1dd818b
Fix the node's order issue when the content of graph is changed ( #16088 )
...
* Fix the node's sort issue when the graph is changed.
test=develop
* Clean code
test=develop
6 years ago
fab1b54d99
Merge branch 'add-communicator' of ssh://github.com/jacquesqiao/Paddle into add-async-ssa-graph-executor-communicator
6 years ago
2eff3e26b6
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add_static_model_load_for_trt
6 years ago
06a088a199
fix comments and fix cpplint
...
test=develop
6 years ago
72253391b6
Add MKL-DNN placement pass tester
...
test=develop
6 years ago
49f2f4f91d
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add-communicator
6 years ago
6a2bc9a275
Add Conv Residual Connection UT for Projection
...
test=develop
6 years ago
a6e3cd5eb7
Merge pull request #15425 from panyx0718/api
...
Pass graph to parallel executor instead of program
6 years ago
b8491bfd4e
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into add-communicator
6 years ago
0362ef75f4
fix
...
test=develop
6 years ago
12a0e2ed9d
polish codes
...
test=develop
6 years ago
19d78f6797
polish
...
test=develop
6 years ago
32d5a16036
resolve conflicts
...
test=develop
6 years ago
c4faf36e7a
MKL-DNN: Add test for conv bias fuse pass ( #15824 )
...
* MKL-DNN: Add test for conv bias fuse pass
test=develop
* Remove const cast from Conv Bias Pass Test
* Add conv with bias test case for conv+bias fuse ut
test=develop
6 years ago
26e32e095a
allow compiler to use graph
...
test=develop
6 years ago
0b926114c0
add override to ApplyImpl
...
and #pragma once in edited headers
add #include<string> in edited headers
test=develop
6 years ago
6019054cdd
Merge pull request #15716 from Yancey1989/refine_pg
...
Refine ParallelGraph Execution
6 years ago
e1c707fe9c
fix warnings ( #15790 )
...
* fix warnings
test=develop
* fix enforce test
test=develop
6 years ago
4b193db14c
polish code test=develop
6 years ago
642fd68ce0
update by comment test=develop
6 years ago
077d12b939
fix scale cleaner ( #15742 )
6 years ago
ecc12fb430
3. when runing in trt mode, do not allocate memory for parameters in fluid.
...
test=develop
6 years ago
ecdd1166b8
cleanup code test=develop
6 years ago
73005ee00d
cleanup code test=develop
6 years ago
88d3dc949e
Merge branch 'develop' of github.com:PaddlePaddle/Paddle into refine_pg
...
test=develop
6 years ago
f3463ecb6e
refine pg execution
6 years ago
f9ac88e1a0
Merge pull request #15694 from liuwei1031/fix_security_issue
...
Fix security issue
6 years ago
e49706c80e
Merge pull request #15659 from GBuella/add_to_string
...
Tests - add some missing to_string calls
6 years ago
b1f97a6fa9
fix security issue 27, 38 test=develop
6 years ago
da9c94da33
Clang build fixes ( #15628 )
...
* Remove some superfluous std::move calls
The std:move triggered a build error (with -Werror):
```
[ 9%] Building CXX object paddle/fluid/memory/allocation/CMakeFiles/allocator_facade.dir/allocator_facade.cc.o
/home/tej/code/gbuella_paddle/paddle/fluid/memory/allocation/allocator_facade.cc:86:29: error: moving a temporary object prevents copy elision [-Werror,-Wpessimizing-move]
[this] { return std::move(CreateAllocatorWithChunk()); }, capacity);
^
/home/tej/code/gbuella_paddle/paddle/fluid/memory/allocation/allocator_facade.cc:86:29: note: remove std::move call here
[this] { return std::move(CreateAllocatorWithChunk()); }, capacity);
^~~~~~~~~~ ~
1 error generated.
```
See: https://reviews.llvm.org/D7633
* Remove a superfluous lambda capture from framework/operator.h
```
[ 10%] Building CXX object paddle/fluid/platform/CMakeFiles/device_context.dir/init.cc.o
In file included from /home/tej/code/gbuella_paddle/paddle/fluid/platform/init.cc:19:
/home/tej/code/gbuella_paddle/paddle/fluid/framework/operator.h:229:21: error: lambda capture 'this' is not used [-Werror,-Wunused-lambda-capture]
[this](Variable* var) { return var; });
^~~~
1 error generated.
```
Changing it to `return it->second;`, as is in the function below.
* Rethrow an exception (instead of copying it)
```
[ 11%] Building CXX object paddle/fluid/framework/CMakeFiles/operator.dir/operator.cc.o
/home/tej/code/gbuella_paddle/paddle/fluid/framework/operator.cc:191:13: error: local variable 'exception' will be copied despite being thrown by name [-Werror,-Wreturn-std-move]
throw exception;
^~~~~~~~~
/home/tej/code/gbuella_paddle/paddle/fluid/framework/operator.cc:191:13: note: call 'std::move' explicitly to avoid copying
throw exception;
^~~~~~~~~
std::move(exception)
```
See https://reviews.llvm.org/D43322 for an explanation of this diagnostic message.
* Remove an unused variable
```
/home/tej/code/gbuella_paddle/paddle/fluid/framework/operator.cc:884:16: error: private field 'scope_' is not used [-Werror,-Wunused-private-field]
const Scope& scope_;
^
```
* struct ComputationOpHandle -> class ComputationOpHandle
```
[ 13%] Building CXX object paddle/fluid/framework/details/CMakeFiles/memory_early_delete_pass.dir/memory_early_delete_pass.cc.o
In file included from /home/tej/code/gbuella_paddle/paddle/fluid/framework/details/memory_early_delete_pass.cc:21:
/home/tej/code/gbuella_paddle/paddle/fluid/framework/details/reference_count_pass_helper.h:30:1: error: class 'ComputationOpHandle' was previously declared as a struct; this is valid, but may result in linker errors under the Microsoft C++ ABI [-Werror,-Wmismatched-tags]
class ComputationOpHandle;
^
/home/tej/code/gbuella_paddle/paddle/fluid/framework/details/computation_op_handle.h:29:8: note: previous use is here
struct ComputationOpHandle : public OpHandleBase {
^
/home/tej/code/gbuella_paddle/paddle/fluid/framework/details/reference_count_pass_helper.h:30:1: note: did you mean struct here?
class ComputationOpHandle;
^~~~~
struct
1 error generated.
```
* Fix name() methods under fluid/operators
```
In file included from /home/tej/code/gbuella_paddle/paddle/fluid/operators/jit/gen/act.cc:15:
In file included from /home/tej/code/gbuella_paddle/paddle/fluid/operators/jit/gen/act.h:19:
/home/tej/code/gbuella_paddle/paddle/fluid/operators/jit/gen/jitcode.h:71:23: error: 'name' overrides a member function but is not marked 'override' [-Werror,-Winconsistent-missing-override]
virtual const char* name() const = 0;
^
/home/tej/code/gbuella_paddle/paddle/fluid/operators/jit/gen_base.h:31:23: note: overridden virtual function is here
virtual const char* name() const = 0;
^
```
test=develop
6 years ago
1905f1a108
bug fix && test=develop
7 years ago
5cf0092825
add more log and fix test_dist_base in multi_batch_merge_pass
7 years ago
4975a9050a
Tests - add some missing to_string calls
...
```
/home/tej/code/gbuella_paddle/paddle/fluid/framework/ir/seqpool_concat_fuse_pass_tester.cc:167:40: error: adding 'int' to a string does not append to the string [-Werror,-Wstring-plus-int]
std::string prefix = "seqpool_op_" + i;
~~~~~~~~~~~~~~^~~
/home/tej/code/gbuella_paddle/paddle/fluid/framework/ir/seqpool_concat_fuse_pass_tester.cc:167:40: note: use array indexing to silence this warning
std::string prefix = "seqpool_op_" + i;
^
& [ ]
1 error generated.
```
test=develop
7 years ago
238ef94702
fix the build issue on gpu mode for win
...
test=develop
7 years ago
74bc55c2a6
Merge pull request #14975 from dzhwinter/ir_inplace_pass
...
Ir inplace pass
7 years ago
dc5e25fc7f
remove dot marked node ( #15606 )
7 years ago
9c9ad7d40b
Merge remote-tracking branch 'origin/develop' into feature/ir_inplace_pass
...
test=develop
7 years ago
897789b16e
fix save_inferece_model bug ( #15365 )
7 years ago
6f9904e99a
rerun windows ci. test=develop
7 years ago
a52be7c081
refine build strategy. test=develop
7 years ago
b1bdcd4de8
Make separate folders for mkldnn codes
...
test=develop
7 years ago
ee3aae56cd
merge develop branch. test=develop
7 years ago
d6d3e6afe2
add more skip strategy
7 years ago
0779e35544
fix two bug:
...
1. graph and program_desc alignment
2. trt stream
test=develop
7 years ago
3ce6172052
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into graph_quantization
7 years ago
8f3b252392
squash commits. test=develop
7 years ago
885c4e57ab
fea/infer memory optim2 ( #14953 )
7 years ago
9f8f0fc2d3
Memory optimization of depthwise conv op and group norm op ( #15313 )
...
* mem opt
* test=develop
* test=develop
* test=develop
* test=develop
* test=develop
* test=develop
* test=develop
* refine code test=develop
* refine code test=develop
* refine code test=develop
* refine code test=develop
* refine with cub test=develop
* fix mkldnn test && remove comments && test=develop
* polish code && test=develop
* add only_forward test && test=develop
7 years ago
451896fce4
init quantization.
7 years ago
d618e48309
fix fuse square mat order and refine test
...
test=develop
7 years ago
a5d2a6d1ad
add fuse pass of sequared mat sub fusion
7 years ago
ca6fdc6e33
refine and fix test
...
test=develop
7 years ago
a89296ac1f
add repeated fc relu pass
7 years ago
98e85f3735
add_transpose_flatten_concat_fuse ( #15121 )
7 years ago
c3a9f3c4b2
fix typo and refine
...
test=develop
7 years ago
ab9c4b2a9f
refine seqpool concat pass and remove unused nodes
...
test=develop
7 years ago
ce909664d8
Merge remote-tracking branch 'ups/develop' into refine/seqpool/feed
7 years ago
fb63cd89d4
Add python ir graph API ( #14917 )
7 years ago
a0a27bd240
add seqpool concat fuse pass tester
...
test=develop
7 years ago
8e086a8521
follow comment and fix typo
...
test=develop
7 years ago
72d2a1801e
add seqpool concat fuse pass
...
test=develop
7 years ago
b76695418a
Polish log
...
test=develop
7 years ago
4bfa110fd8
Add no lock optimize pass
...
test=develop
7 years ago
e26cced7cc
refine batch merge pass ( #14777 )
...
* refine batch merge pass
* refine batch merge pass test=develop
7 years ago
4048cfa9da
Merge pull request #15048 from NHZlX/add_affine_channel_fuse
...
Add conv+ affine channel fuse pass
7 years ago
fe8495a758
[WIP] Refine MultiDevSSAGraph ( #15040 )
...
* refine parallel_exe
test=develop
* rename shared_var_device
* code refine
* add test_weight_decay
* remove Sort
test=develop
* Add SortForReduce
test=develop
* code refine
test=develop
* follow comment
test=develop
7 years ago
a6aa8ea771
faster rcnn input is presistable. (fix it in paddle-trt)
...
test=develop
7 years ago
956cf92145
Fix conv_elementwise_add2_act pass
...
test=develop
7 years ago
ce3782c193
add affine_channel fuse.
...
fix conv+elemenwise fuse bug.
7 years ago
e9eee0de6a
MKLDNN residual connection fuse: fixing accuracy problem ( #14874 )
...
* MKLDNN residual connection fuse: conv op reused
test=develop
* MKLDNN residual connection fuse: added prints for checking fuse
test=develop
* MKLDNN residual connection fuse: add more prints
test=develop
* MKLDNN residual connection fuse: add hash function. test=develop
* MKLDNN residual connection fuse: add hash to elementwise_add
test=develop
* MKLDNN residual connection fuse: add more hashes. test=develop
* MKLDNN residual connection fuse: added hashes to relu
test=develop
* MKLDNN residual connection fuse: do not fuse when fuse_relu is on
* MKLDNN residual connection fuse: check if fuse_relu attribute is set
test=develop
* MKLDNN residual connection fuse: comment out some printouts
* MKLDNN residual connection fuse: remove unused functions in the pass code
* MKLDNN residual connection fuse: delete commented hashes and printouts
* MKLDNN residual connection fuse: remove unnecessary includes. test=develop
7 years ago
a9fb34fad8
Merge pull request #14903 from NHZlX/add_conv_elementwise_pass
...
Add conv + elementwiseAdd pass
7 years ago
7cd24b1318
add ir memory optimize. ( #14530 )
...
* follow comments. test=develop
* Fix typo
* fix compile error. test=develop
* merge develop branch. test=develop
* Remove set_equal
* Polish code
* Delete unused functions
test=develop
* polish code. test=develop
* follow comment
* polish code.
* fix windows compile error. test=develop
* fix op handle.
* rerun ci. test=develop
* rerun ci. test=develop
* rerun macci. test=develop
* polish code. test=develop
* rewrite sort code. test=develop
* remove unused code. test=develop
* fix tests. test=develop
* fix conflict. test=develop
* follow comment. test=develop
* merge develop branch. test=develop
* fix tests. test=develop
* remove ToTypeIndex. test=develop
* rerun ci. test=develop
7 years ago
c0c9fcd9c7
add source file
...
test=develop
7 years ago
4e4a777243
add conv+elementwiseadd pass
...
test=develop
7 years ago
a985949be9
Fea/fuse conv elementwise add fuse ( #14669 )
7 years ago
7923042365
merge develop
...
test=develop
7 years ago
3243b45bb2
Merge pull request #14651 from luotao1/has_attr
...
add Opdesc's HasProtoAttr
7 years ago
067ed70f2d
add HasProtoAttr function in op_desc.h, clean node.h
...
test=develop
7 years ago
748549b2e3
Revert "Merge pull request #14798 from PaddlePaddle/revert-14786-revert-14782-revert-14398-imperative"
...
This reverts commit b1d3a1c8b4f1fb64b17f
7 years ago
943ad4781f
One possible solution to add flexibility for mkldnn placement pass ( #14768 )
...
* Choose to turn on use_mkldnn attribute v1
* Fix mkldnn_op empty bug
* format change test=develop
* fix ci test=develop
* fix ci test and add test in dam test=develop
* add example to dam compare test test=develop
* review changes test=develop
7 years ago
c049fa7cf7
Revert "Revert "Revert "Imperative"""
7 years ago
c83d5b7a16
Merge pull request #14709 from yihuaxu/develop_4f71a6ee2_conv3d_bias_fusion_mkldnn_impl
...
Implement the fusion of convolution 3D and bias for mkldnn
7 years ago
3821fc3950
Merge branch 'develop' into develop_4f71a6ee2_conv3d_bias_fusion_mkldnn_impl
...
test=develop
7 years ago
240d974ac5
Clean Code
...
test=develop
7 years ago
2538ef64f1
Revert "Revert "Imperative""
7 years ago
6217f42ab7
Revert "Imperative"
7 years ago
b52f5d2870
Merge pull request #14398 from panyx0718/imperative
...
Imperative
7 years ago
0591ba96ec
fix hack
...
test=develop
7 years ago
64ad051b9a
merge develop
...
test=develop
7 years ago
64e261c6cd
Implement the fusion of convolution and bias for mkldnn
...
(test=develop)
7 years ago
8d6984eb9b
change OpHasAttr to RuntimeHasAttr, add some comments
...
test=develop
7 years ago
aeb74af54c
allow operator to run imperatively
7 years ago
096673f675
refactor eager deletion
...
test=develop
7 years ago
c856ac8721
add OpHasAttr in node.h, update is_test_pass and mkldnn_placement_pass
...
test=develop
7 years ago
be04d99fe4
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into revert_vlog
...
test=develop
7 years ago
53433d7f2e
Revert the changes of VLOG
...
test=develop
7 years ago
7c8c9dc9bf
fix unit test cases
7 years ago
a94a7355f0
Refine the GraphNum check ( #14144 )
...
* refine GraphCheck
test=develop
* fix ci fail
test=develop
7 years ago
53da846d1e
MKLDNN residual connections fuse pass: initial implementation of fusion for projection pass
...
test=develop
7 years ago
dbc4fcd722
MKLDNN residual connections fuse pass: unit tests enabled and added
7 years ago
4224089354
MKLDNN residual connections fuse pass: Maybe removed and boost::optional used where it makes sense
7 years ago
86fd3b32be
MKLDNN residual connections fuse pass: counting statistics added to the pass
7 years ago
ee6f778beb
MKLDNN residual connections fuse pass: further refactoring
7 years ago
7423748e37
MKLDNN residual connections fuse pass:
...
* implements reachability check between identity node and non-identity argument to elementwise_add
* implements handling identity node as x and as y argument to elementwise_add
7 years ago
8a1eeec579
add mkldnn prop_kind phase for inference-only case to pooling and activations ( #14278 )
...
* add is_test to pooling and activations
add prop_kind support for layers activation. conv and pooling
add a pass that sets is_test to true
add transpiler version of is_test pass
test=develop
* patch test and pass
test=develop
* add pass to analyzer.h
test=develop
* add is_test attr description & pass only on mkldnn
in:
activation_op.cc
batch_norm_op.cc
conv_op.cc
dropout_op.cc
lrn_op.cc
pool_op.cc
sequence_pool_op.cc
softmax_op.cc
* fix is_test handling for activation pool and conv
* change description of is_test for all layers again
* remove GetAttr(use_mkldnn) from pass
* rename correct_mkldnn_test_phase to is_test
and remove dependency on MKLDNN
test=develop
* review fix magic number
* two if(..)s into one
* Check is_test once and pass mkldnn forward prop kind
* dereference shared_ptr with * (without get())
test=develop
* add is_test_pass back
test=develop
7 years ago
5ef123c778
Merge branch 'develop' into dam_fc
7 years ago
980a6753a8
fix typo to pass the ci
...
test=develop
7 years ago
1a9008c420
code style fix
...
test=develop
7 years ago
8ea13e336a
add in_num_col_dims for fc
7 years ago
08d1dc84a9
fix
7 years ago
30ddc07a7e
Merge remote-tracking branch 'upstream/develop' into windows/build
7 years ago
9f252e0032
Combine Inference Analysis with IR ( #13914 )
7 years ago
6c32945556
Merge pull request #14372 from luotao1/speedup_analysis
...
speedup DetectPatterns
7 years ago
668ae523d2
speedup DetectPatterns
...
test=develop
7 years ago
9a6e239281
fix mac graph detector sort ( #14356 )
7 years ago
dc339b78d7
fix code style
7 years ago
ca60e1d34d
Merge remote-tracking branch 'upstream/develop' into windows/build
7 years ago
d08334011a
fix merge issue
7 years ago
52f7644f53
Merge remote-tracking branch 'upstream/develop' into windows/build
7 years ago
abe209234f
Exhaustive search for cuDNN conv. ( #14286 )
...
* exhaustive search for cuDNN conv.
* Refine code and add unit testing.
* Fix model load in fluid/inference and unit testing in conv2d
* Follow comments.
* Fix compiling test=develop
7 years ago
87450b9ad4
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_vlog
...
test=develop
7 years ago
4ffa92d4f0
Merge branch 'develop' into windows/build
7 years ago
813e54efbd
Merge pull request #14328 from PaddlePaddle/revert-14046-windows/debug
...
Revert "cherry picked windows patches."
7 years ago
3db9fad764
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_vlog
...
test=develop
7 years ago
ff6c809bfc
Merge pull request #14251 from panyx0718/fix
...
Make OpHandle/VarHandle and ir::Node works cleaner
7 years ago
ba8b5619a3
Revert "cherry picked windows patches."
7 years ago
fcc0452c8b
Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into fix_vlog
...
test=develop
7 years ago
0c3227a523
Change the origin VLOG level to 10 times
...
Fix code to support cpplint syntax check
test=develop
7 years ago
59c66532e7
add more logs and comments
...
test=develop
7 years ago
869487a2b7
Merge remote-tracking branch 'origin/develop' into windows/build
7 years ago
2835e04409
merge develop branch. test=develop
7 years ago
db8c52da5e
Revert " Exhaustive search for cuDNN conv. ( #14043 )"
...
This reverts commit ce7d9b0799
7 years ago
ce7d9b0799
Exhaustive search for cuDNN conv. ( #14043 )
...
* exhaustive search for cuDNN conv.
* Refine code and add unit testing.
* Clean code
* Fix model load in fluid/inference and unit testing in conv2d
* Follow comments.
7 years ago
25123a3b7e
add tests
...
test=develop
7 years ago
8c11d3fed6
clean up
7 years ago
adf5615e54
clean kGraphOp
...
test=develop
7 years ago
fb576cb5cb
allow to compare type
...
test=develop
7 years ago
ead94bfc6c
fix destructor
...
test=develop
7 years ago
2e14999942
clean1
...
test=develop
7 years ago
9d67c1fb69
cpu build support
7 years ago
c2d70fca30
fix to only check block 0
...
test=develop
7 years ago
aaeedd0ff3
make it warn
...
test=develop
7 years ago
ddd2225b56
add more debug info.
...
test=develop
7 years ago
d2a56f7909
Merge pull request #14159 from sfraczek/sfraczek/depthwise-conv-mkldnn-pass
...
add depthwise conv mkldnn pass
7 years ago
eb2f7ed21b
refine tests. test=develop
7 years ago
08d22cf7e1
Merge pull request #14091 from panyx0718/fix2
...
add program check
7 years ago
f76fee644c
fix graph pattern detector ( #14186 )
7 years ago
4e2aaf01bc
add depthwise conv mkldnn pass
...
added depthwise conv mkldnn pass which for MKLDNN changes depthwise_conv operator to conv operator because for mkldnn this is the same api
test=develop
7 years ago
a943134a97
fix a few more tests
...
test=develop
7 years ago
5839e3236b
add program check
...
test=develop
7 years ago
ebfe5a02b3
merge develop branch
7 years ago
26200f2e42
[1.1] [project] train imagenet using large batch size ( #13766 )
...
* fix nccl2 lars dist support
* put lars in momentum op
* add tests lars
* fix ci
* fix cpu kernel
* soft warning
* remove lars in test_recognize_digits.py
* move to another op
* add file
* update api.spec test=develop
* update test=develop
* fix api.spec test=develop
* wip
* wip, finish grad merge ops
* wip, finish graph build
* wip test running
* work on 1 gpu
* workable version
* update
* fix tests
* fuse broadcast op
* fix compile failed
* refine
* add batch merge test mnist
* fix CI test=develop
* fix build
* use independent bn params for batch merge test=develop
* update api.spec
* follow comments and for test
* wip
* refine tests test=develop
* follow comments test=develop
* remove startup bn modify test=develop
* follow comments test=develop
* fix merge test=develop
7 years ago
0a80f06ec4
Merge pull request #14086 from panyx0718/fix6
...
delete unused codes.
7 years ago
9da9b1926b
[1.1] fix graph num hang ( #14072 )
...
* fix graph num hang test=develop
* re-enable tests test=develop
* re-enable graph num check test=develop
* fix multi device pass role check test=develop
7 years ago
bba0c4a9f2
delete unused codes.
...
test=develop
7 years ago
78cf76a1ca
fix linux compile
7 years ago
316bc9bfc9
fix typo and warning in analyzer_resnet50_test
...
test=develop
7 years ago
42aa1d409d
Merge pull request #13485 from tpatejko/tpatejko/capi-resnet-conv-elementwise-fusion
...
MKLDNN conv+elementwise_add fusion for residual connections in Resnet
7 years ago
ce2464fd98
MKLDNN conv + elementwise_add fusion: UT for missing bias added. UTs refactored. Some minor changes in the pass
7 years ago
4e72ab411e
MKLDNN conv + elementwise_add fusion: fix for crash when bias is not present
7 years ago
415b261555
MKLDNN conv + elementwise_add fusion: fusion options added
7 years ago
0fe3079c46
MKLDNN conv + elementwise_add fusion: fix for order of parameters in elementwise_add in resnet50
...
test=develop
7 years ago
b73b868366
MKLDNN conv + elementwise_add fusion: bias in tests made persistent.
...
test=develop
7 years ago
a1fa203287
MKLDNN conv + elementwise_add fusion: name of the pass reused with name_scope_
7 years ago
2c43419db1
MKLDNN conv + elementwise_add fusion: comment explaining CorrectGraphEdges added
7 years ago
8fb29b2ca9
MKLDNN conv + elementwise_add fusion: new nodes marked as input or output
...
test=develop
7 years ago
cc1c8e37c1
MKLDNN conv + elementwise_add fusion: attributes in new conv op copied from old op
7 years ago
a27a8c5da8
MKLDNN conv + elementwise_add fusion: bias in test marked as persistable
7 years ago
af8c71317c
MKLDNN conv + elementwise_add fusion: CorrectGraphEdges refactored
7 years ago
3e033087f1
MKLDNN conv + elementwise_add fusion: LinkNodes function removed and
...
macro used.
test=develop
7 years ago
4be45af1cc
MKLDNN conv + elementwise_add fusion: skip connection attribute renamed. Comments about patterns added.
...
test=develop
7 years ago
9a335e0277
MKLDNN conv + elementwise_add fusion: changed a name of a formal argument in ElementwiseAdd pattern
7 years ago
fb7a50b230
MKLDNN conv + elementwise_add fusion: removed commented code. Internal functions marked as static.
...
test=develop
7 years ago
efd76614fb
MKLDNN conv + elementwise_add fusion: implementation changed to conform with Paddle API
7 years ago
347bf90412
MKLDNN conv + elementwise_add fusion: bias is also handled
7 years ago
bf95ac36a7
MKLDNN conv + elementwise_add fusion: further reformatting
7 years ago
cbe122ae2e
MKLDNN conv + elementwise_add fusion: correcting formatting
7 years ago
2a251bbf27
MKLDNN conv + elementwise_add fusion: some refactoring: consts, function calls instead of constant values
7 years ago
b8e54ab5cc
MKLDNN conv + elementwise_add fusion: parameter name changed to ResidualData
7 years ago
27573ece03
MKLDNN conv + elementwise_add fusion: trailing spaces removed
7 years ago
7f5c8a95e8
MKLDNN conv + elementwise_add fusion: arguments are replaced for many parameters in operator
7 years ago
5996bd39e8
MKLDNN conv + elementwise_add fusion: graph is corrected based on actual argument name, not formal argument name
7 years ago
41f3d78fdf
MKLDNN conv + elementwise_add fusion: output and elemwise param share data in conv primitive. Output is properly allocated
7 years ago
07a62ddc08
MKLDNN conv + elementwise_add fusion: inputs in pass modified. Support for new conv parameter. UTs corrected
7 years ago
42f569fdfd
MKLDNN conv + elementwise_add fusion: use_mkldnn attribute added
7 years ago
441d3a4726
MKLDNN conv + elementwise_add: added some refactoring in the pass
7 years ago
38b7b34b1c
MKLDNN conv + elementwise_add fusion: added reachability tests, inputs and outputs in graph nodes are transformed
7 years ago
16eaaf3fbe
MKLDNN conv + elementwise_add fusion: added one more UT, found and corrected bugs in pass
7 years ago
604bad08bc
MKLDNN conv + elementwise_add fusion: implementation of patterns refarctored, applied to graph. UTs added
7 years ago
9ce343f868
MKLDNN conv + elementwise_add fusion: initial implementation of patterns
7 years ago
40f8456a4f
refine fuse pattern and attr
...
test=develop
7 years ago
cbbacb2534
Merge remote-tracking branch 'ups/develop' into fea/fusion_seqconv_add
...
test=develop
7 years ago
603ba5e01d
add seqconv eltadd relu pass
7 years ago
c504a5a1b7
Adjust Conv+bias to placement pass
...
test=develop
7 years ago
d7509d63f1
Conv+Bias: Support non-null bias
...
test=develop
7 years ago
582f59c190
Conv+Bias fuse
7 years ago
c3b70aece9
Add MKL-DNN placement pass ( #13958 )
...
* add MKL-DNN placement pass
This patch also refactors conv+bn (includes changes from PR
https://github.com/PaddlePaddle/Paddle/pull/13926 )
updated to use the mkldnn-placement-pass.
test=develop
* remove redundant pass list
* add comment on the default first pass
* fix test for conv+relu mkldnn fuse
7 years ago
7b11162ab5
Merge pull request #13949 from PaddlePaddle/wojtuss/unique-patterns-request-comment
...
Added comment with request for enhancement
7 years ago
55fd136ab0
Added comment with request for enhancement
...
This adds a `TODO` comment according to https://github.com/PaddlePaddle/Paddle/issues/13550#issuecomment-430133585
test=develop
7 years ago
e47f4186ae
fix some compiler warning
7 years ago
e41a3fcd68
fix update to develop hang problem.
7 years ago
50c5e9b0c6
reshape_2d used from ddim.h
...
test=develop
7 years ago
3ae9645084
compile in linux
7 years ago
55d6950a1a
rewrite conv_bn fuse pass to eigen
...
test=develop
7 years ago
9b11a17502
Revert "[MKLDNN] Pass: Fuse Conv + Bias"
7 years ago
ce248a15d9
Merge pull request #13368 from Sand3r-/mgallus/conv-bias-pass
...
[MKLDNN] Pass: Fuse Conv + Bias
7 years ago
40b17be4b0
Pass: Fuse Conv + Bias
...
test=develop
7 years ago
3fcca40909
eigen sqrt fix and change 1e-5 to epsilon
...
test=develop
7 years ago
78f98294c2
conv bn fuse pass
...
review fix
review from hshen14 fix
test=develop
fix error in broadcast and code cleanup
rename bias -> eltwise and added macro to shorten code
formatting
7 years ago
a89afd4c22
Merge pull request #13685 from luotao1/naive_cmake
...
update libpaddle_fluid.a/so
7 years ago
a989a4e7c2
refine paddle_inference_helper.h
7 years ago
ede4b230be
Merge pull request #13553 from jczaja/prv-fused_embedding_fc_lstm_op
...
Adding fused_embedding_fc_lstm op
7 years ago
e202f33aa9
- Yet another clarification to comment
...
test=develop
7 years ago
1df69f7c9d
- Fix to comment
...
test=develop
7 years ago
c8744d118d
fea/infer executor and concurrency performance issue bug fix ( #13451 )
...
- add naive executor
- fix concurrency performance issue
7 years ago
5175b3cb2b
Add GraphChecker ( #13580 )
...
* add GraphNum
test=develop
* add graph number check in parallelExecutor
test=develop
* fix transformer_model bug
test=develop
* fix graph num
7 years ago
7cd2761736
Merge pull request #13416 from panyx0718/ir
...
PassBuilder
7 years ago
d5114c60b0
- Reviewers suggesstions to fused_embedding_fc_lstm_op
7 years ago
7ab5626dee
- Added initial pass for embedding-fc-lstm
...
- Added draft of new operator
- Added fused embedding fc lstm files
- First time embedding_fc_lstm_fuse_pass was invoked in
test_text_classification
- Added Embedding pattern
- Not crashing
- Enabled draft of embedding_fc_lstm pass (does it job)
- First working (Seqcompute only) version
- Removed diagnostic comment
- First enabling of BatchCompute
- Disabling pass for embedding with is_sparse and is_distributed
- Cosmetics
- Style
- Style
7 years ago
9e8d372ff4
hide attention lstm fuse ( #13615 )
7 years ago
6938e6cf06
Merge pull request #13603 from tensor-tang/refine/peephole
...
refine peephole
7 years ago
16e73e0d28
hide operator API ( #12543 )
...
* hide operator API
* update
* update api.spec
* fix merge
* fix test
7 years ago
209e9c3db1
refine peephole
...
test=develop
7 years ago
e5d1bd1e93
remove unused variable nodes2delete
7 years ago
a49aa4dac9
make bias unnecessary for ConvRelu fuse
7 years ago
493ef0c8df
do not remove conv node just rewire the output
7 years ago
667b661786
updated the test
7 years ago
c2b3838cf5
add some comments
7 years ago
b43e49fa31
fix
7 years ago
36c2a9af27
pass builder allow cutomize pass in python.
7 years ago
d402234ba8
Feature/op_fuse_pass ( #12440 )
...
* Add Preface
* Add demo code
* Save file
* Refine code
* seems can work
* use elementwise strategy
* Use ElementwiseComputeEx
* Add comments
* extract functions from operator
* Refine code
* Follow comment
* code refine
* add op_fuse pass
* add backward
* code refine
* use TopologySortOperations
* follow comments
* refine IsFusible
* code enhance
* fix op_fusion_pass
* refine code
* refine fuse_elemwise_act_op
* adjust the input and output
* refine logic
* add intermediate_edge
* disable inplace
* follow comments
* refine logic
* follow comments
* Remove the removable IntermediateOut
* change strategy
* code refine
* enable fuse backward
* code refine
* code refine
* rename unit test
* follow comments
7 years ago
23ba766252
Merge pull request #13475 from panyx0718/ir5
...
avoid creating dangling ir::Node.
7 years ago
0bd7a67eaf
avoid creating dangling ir::Node.
...
Node should be created by Graph::CreateXXX so that
they are managed by graph.
7 years ago
d0b2453ecd
merge develop
7 years ago
minqiyang
luotao1
Wojciech Uss
qingqing01
Zhen Wang
Yihua Xu
Qiao Longfei
nhzlx
Krzysztof Binias
Michal Gallus
Xin Pan
Michał Gallus
Sylwester Fraczek
tensor-tang
Yancey1989
Yan Chunwei
dzhwinter
liuwei1031
Gabor Buella
Dun Liang
peizhilin
dzhwinter
flame
Wu Yi
chengduo
hjchen2
Tomasz Patejko
sneaxiy
bingyanghuang
Yihua Xu
wanghaoshuang
Jacek Czaja