Commit Graph

3082 Commits (02e17396c24f0deb11826e37a579a69dc41ca382)

Author SHA1 Message Date
nhzlx 73b47df1f4 Merge branch 'develop' of https://github.com/paddlepaddle/paddle into add_affine_channel_fuse
6 years ago
nhzlx ce3782c193 add affine_channel fuse.
6 years ago
Qiyang Min aba1f9b06e
Merge pull request #14891 from velconia/accelerate_adam
6 years ago
tensor-tang 05d1121b22
Merge pull request #14802 from mozga-intel/mozga-intel/fill_constant_operator_ngraph
6 years ago
tensor-tang 9d4f1d468a
Merge pull request #14804 from mozga-intel/mozga-intel/top_k_operator_ngraph
6 years ago
tensor-tang 8a6ac4dba7
Merge pull request #14973 from xiaolil1/dequantize
6 years ago
tensor-tang f0e02a65ed
Merge pull request #14974 from xiaolil1/quantize
6 years ago
qingqing01 51a9fca323
Async memory copy (#15013)
6 years ago
xiaoli.liu@intel.com 869d444b92 Fix comments misunderstanding
6 years ago
xiaoli.liu@intel.com d83d0f33fd extract templated function
6 years ago
minqiyang 8b6b0da062 Use adam_update
6 years ago
minqiyang f4e7a47381 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into accelerate_adam
6 years ago
tensor-tang 641313ea77
Merge pull request #15000 from tensor-tang/doc/en/jit
6 years ago
wopeizl b117a5f208
Merge pull request #14931 from wopeizl/windows/mkl
6 years ago
tensor-tang 68ab16444a add eng doc of jit kernel and follow comments
6 years ago
chengduo 79bd6dfa18
[Feature] Add Temporary Allocator (#14875)
6 years ago
peizhilin 2e35290ffa fix build issue
6 years ago
Xin Pan 484c24b756
Merge pull request #14996 from tensor-tang/fix/deps
6 years ago
tensor-tang e99e4e994d fix deps of jit benchmark
6 years ago
Tao Luo d4f384767a
Merge pull request #14987 from jczaja/prv-transpose-mkldnn-grad
6 years ago
peizhilin 201283f95c fix code style
6 years ago
peizhilin 9e60c58666 Merge remote-tracking branch 'upstream/develop' into windows/mkl
6 years ago
tensor-tang 693e5e65ce
Merge pull request #14958 from tensor-tang/refine/jit
6 years ago
Zeng Jinle 95cbe07c40
Merge pull request #14836 from sneaxiy/feature/py_func
6 years ago
mozga-intel 969f669493 Static specifier was removed
6 years ago
mozga-intel 7048caf9a0 Enable top_k operator for a ngraph
6 years ago
mozga-intel 394635b2ee Static specifier was removed
6 years ago
mozga-intel ecfa68ecaa Enable fill_constant operator for a ngraph
6 years ago
Jacek Czaja d05094eec3 - Added transpose/transpose2 MKLDNN grad ops
6 years ago
Tao Luo 55af116861
Merge pull request #14966 from jczaja/prv-transpose-mkldnn-reuse-PR
6 years ago
Xin Pan 2537ac512e
Merge pull request #14983 from typhoonzero/revert_14850
6 years ago
chengduo a015a8a39d
Refine merge_selected_rows Doc (#14748)
6 years ago
typhoonzero da87f7a698 Revert "[Feature] Fp16 training for resnet50 (#14850)"
6 years ago
tensor-tang 1aaec571c2 fix enum style
6 years ago
sneaxiy 490eb9061f polish infer shape of py_func op
6 years ago
tensor-tang facfecbdd0 follow comment: reuse time function and change to upper case
6 years ago
tensor-tang f5532877f7 Merge remote-tracking branch 'ups/develop' into refine/jit
6 years ago
Yihua Xu 3babc80160 Fix the regression issue and add the group unitest for conv2d (#14932)
6 years ago
Xin Pan a872eb90c2
Merge pull request #14959 from panyx0718/clean2
6 years ago
minqiyang 6467eda534 Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into accelerate_adam
6 years ago
Wu Yi 3d750f9c5a
[Feature] Fp16 training for resnet50 (#14850)
6 years ago
Wang Guibao e7675dd542
Avoid calling VLOG() when program exiting (#14937)
6 years ago
xiaoli.liu@intel.com 3146629485 Add Dequantize OP
6 years ago
xiaoli.liu@intel.com 019dbf7f5f Add Quantize OP
6 years ago
tensor-tang a369c8025b fix typo of mix impl of lstm
6 years ago
tensor-tang d651a91138 fix build on win, fix use condition of crf decoding and layer norm and
6 years ago
tensor-tang d53c4756ad clean code and remove unused files
6 years ago
tensor-tang 95fb31285c Merge remote-tracking branch 'ups/develop' into refine/jit
6 years ago
Xin Pan cf3a07e8f8
Merge pull request #14878 from panyx0718/imperative
6 years ago
peizhilin 9f55f1ff50 use the platform api to decide the specific instruction support or not
6 years ago