Commit Graph

322 Commits (0ed63b2108fdbfb683140765dd5a378697593659)

Author SHA1 Message Date
nhzlx ecc12fb430 3. when runing in trt mode, do not allocate memory for parameters in fluid.
6 years ago
peizhilin 238ef94702 fix the build issue on gpu mode for win
6 years ago
Xin Pan 74bc55c2a6
Merge pull request #14975 from dzhwinter/ir_inplace_pass
6 years ago
Yan Chunwei dc5e25fc7f
remove dot marked node (#15606)
6 years ago
dzhwinter 9c9ad7d40b Merge remote-tracking branch 'origin/develop' into feature/ir_inplace_pass
6 years ago
Yan Chunwei 897789b16e
fix save_inferece_model bug (#15365)
6 years ago
dzhwinter 6f9904e99a rerun windows ci. test=develop
6 years ago
dzhwinter a52be7c081 refine build strategy. test=develop
6 years ago
Krzysztof Binias b1bdcd4de8 Make separate folders for mkldnn codes
6 years ago
dzhwinter ee3aae56cd merge develop branch. test=develop
6 years ago
dzhwinter d6d3e6afe2 add more skip strategy
6 years ago
nhzlx 0779e35544 fix two bug:
6 years ago
WangZhen 3ce6172052 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into graph_quantization
6 years ago
dzhwinter 8f3b252392 squash commits. test=develop
6 years ago
Yan Chunwei 885c4e57ab
fea/infer memory optim2 (#14953)
6 years ago
Dun 9f8f0fc2d3 Memory optimization of depthwise conv op and group norm op (#15313)
6 years ago
WangZhen 451896fce4 init quantization.
6 years ago
tensor-tang d618e48309 fix fuse square mat order and refine test
6 years ago
tensor-tang a5d2a6d1ad add fuse pass of sequared mat sub fusion
6 years ago
tensor-tang ca6fdc6e33 refine and fix test
6 years ago
tensor-tang a89296ac1f add repeated fc relu pass
6 years ago
Zhaolong Xing 98e85f3735 add_transpose_flatten_concat_fuse (#15121)
6 years ago
tensor-tang c3a9f3c4b2 fix typo and refine
6 years ago
tensor-tang ab9c4b2a9f refine seqpool concat pass and remove unused nodes
6 years ago
tensor-tang ce909664d8 Merge remote-tracking branch 'ups/develop' into refine/seqpool/feed
6 years ago
flame fb63cd89d4
Add python ir graph API (#14917)
6 years ago
tensor-tang a0a27bd240 add seqpool concat fuse pass tester
6 years ago
tensor-tang 8e086a8521 follow comment and fix typo
6 years ago
tensor-tang 72d2a1801e add seqpool concat fuse pass
6 years ago
minqiyang b76695418a Polish log
6 years ago
minqiyang 4bfa110fd8 Add no lock optimize pass
6 years ago
Wu Yi e26cced7cc
refine batch merge pass (#14777)
6 years ago
Zhaolong Xing 4048cfa9da
Merge pull request #15048 from NHZlX/add_affine_channel_fuse
6 years ago
chengduo fe8495a758
[WIP] Refine MultiDevSSAGraph (#15040)
6 years ago
nhzlx a6aa8ea771 faster rcnn input is presistable. (fix it in paddle-trt)
6 years ago
hjchen2 956cf92145 Fix conv_elementwise_add2_act pass
6 years ago
nhzlx ce3782c193 add affine_channel fuse.
6 years ago
Tomasz Patejko e9eee0de6a MKLDNN residual connection fuse: fixing accuracy problem (#14874)
6 years ago
Zhaolong Xing a9fb34fad8
Merge pull request #14903 from NHZlX/add_conv_elementwise_pass
6 years ago
dzhwinter 7cd24b1318
add ir memory optimize. (#14530)
6 years ago
nhzlx c0c9fcd9c7 add source file
6 years ago
nhzlx 4e4a777243 add conv+elementwiseadd pass
6 years ago
Yan Chunwei a985949be9
Fea/fuse conv elementwise add fuse (#14669)
6 years ago
sneaxiy 7923042365 merge develop
6 years ago
Tao Luo 3243b45bb2
Merge pull request #14651 from luotao1/has_attr
6 years ago
Tao Luo 067ed70f2d add HasProtoAttr function in op_desc.h, clean node.h
6 years ago
Xin Pan 748549b2e3 Revert "Merge pull request #14798 from PaddlePaddle/revert-14786-revert-14782-revert-14398-imperative"
6 years ago
bingyanghuang 943ad4781f One possible solution to add flexibility for mkldnn placement pass (#14768)
6 years ago
Xin Pan c049fa7cf7
Revert "Revert "Revert "Imperative"""
6 years ago
Tao Luo c83d5b7a16
Merge pull request #14709 from yihuaxu/develop_4f71a6ee2_conv3d_bias_fusion_mkldnn_impl
6 years ago