mindspore-ci-bot
|
471c203443
|
!7712 rename and change lib c++_static to c++_shared
Merge pull request !7712 from yeyunpeng2020/primitive
|
5 years ago |
yeyunpeng
|
7945bbf9f2
|
rename and change lib c++_static to c++_shared
|
5 years ago |
yangruoqi713
|
9e274b6468
|
[MSLITE][Develop] optimize arm cpu int8 op conv dw 3x3, add border assembly
|
5 years ago |
fuzhiye
|
58a73bdd5b
|
remove useless code
|
5 years ago |
mindspore-ci-bot
|
5d39dcda6e
|
!7668 [MSLITE]fix multipe definition of function
Merge pull request !7668 from 张学同/to_merge
|
5 years ago |
zhangxuetong
|
61178eaaf9
|
fix multipe definition of function
|
5 years ago |
mindspore-ci-bot
|
9d9f98768b
|
!7643 [MSLITE][Develop] optimize arm cpu int8 op conv dw 3x3: add assembly arm32
Merge pull request !7643 from yangruoqi713/conv_dw
|
5 years ago |
yangruoqi713
|
b1dbfa643b
|
[MSLITE][Develop] optimize arm cpu int8 op conv dw 3x3: add assembly arm32
|
5 years ago |
mindspore-ci-bot
|
35624350b4
|
!7632 [MS][LITE][CPU]change scheme for int8 conv(non-sdot)
Merge pull request !7632 from fuzhiye/tmp
|
5 years ago |
fuzhiye
|
b43cdc5df9
|
1.change int8 conv which not support sdot scheme (using matmul)
2.free reduandent memory in fp16 kernels
|
5 years ago |
zhanyuan
|
4b810f5ee3
|
Optimize depthwise convolution 3x3 for arm64 platform
|
5 years ago |
yangruoqi713
|
89e83b92d0
|
[MSLITE][Develop] optimize arm cpu int8 op conv dw 3x3: add assembly
|
5 years ago |
mindspore-ci-bot
|
98f3fe665c
|
!7532 fix concat memory leak
Merge pull request !7532 from zhaodezan/master
|
5 years ago |
mindspore-ci-bot
|
7e8dad791a
|
!7526 [MSLITE][Develop] optimize conv dw arm cpu int8 op: add 3x3
Merge pull request !7526 from yangruoqi713/conv_dw
|
5 years ago |
yangruoqi713
|
161ecc4ed0
|
[MSLITE][Develop] optimize conv dw arm cpu int8 op: add 3x3
|
5 years ago |
zhaodezan
|
f362df80c1
|
fix concat memory leak
|
5 years ago |
mindspore-ci-bot
|
d7ea4742f3
|
!7466 [MSLITE] Fix bug of arithmetic operator.
Merge pull request !7466 from wangshaocong/bugfix_master_v2
|
5 years ago |
wang_shaocong
|
7d75ac047f
|
Fix bug of several operators.
|
5 years ago |
fuzhiye
|
3d42727c48
|
optimize int8 max pool
|
5 years ago |
fuzhiye
|
68b8c7e84c
|
replace int8 common gemm with int8 matmul
|
5 years ago |
AGroupofProbiotocs
|
5503d78700
|
fix space_to_batch_int8 padding bug
|
5 years ago |
fuzhiye
|
8549266616
|
optimize int8 sigmoid(logistic)
|
5 years ago |
zhanyuan
|
917adc6baf
|
Fix the bug of int8 matmul weight tensor init
|
5 years ago |
wsc
|
6835d45691
|
Fix bug of leakyRelu_Int8 operator.
|
5 years ago |
mindspore-ci-bot
|
dc1368f09d
|
!6733 scale int8
Merge pull request !6733 from zhaozhenlong/lite/op/scale_int8
|
5 years ago |
fuzhiye
|
e14ffdc8f4
|
optimize conv pre process
|
5 years ago |
zhaodezan
|
33b2735560
|
pad int8 surpport mirror pad.
|
5 years ago |
zhaozhenlong
|
e9ea2de302
|
scale int8 code
|
5 years ago |
mindspore-ci-bot
|
d58116644b
|
!6627 [MSLITE][Develop] support conv_depthwise arm32 int8 weight perchannel
Merge pull request !6627 from yangruoqi713/lite
|
5 years ago |
yangruoqi713
|
53c6862a6f
|
[MSLITE][Develop] support conv_depthwise arm32 int8 weight perchannel
|
5 years ago |
liuzhongkai
|
37567147bd
|
code clean
|
5 years ago |
liuzhongkai
|
59bdc993c8
|
fix arm32 ConvDwInt8Center.S
|
5 years ago |
mindspore-ci-bot
|
6f5be6b876
|
!6369 add inferenceType UINT8
Merge pull request !6369 from cjh9368/quant_same_datatype
|
5 years ago |
cjh9368
|
a161a3375a
|
add inferenceType UINT8 and recitify converter params
|
5 years ago |
mindspore-ci-bot
|
0e60ed8927
|
!6351 [MS][LITE][CPU]fix int8 function call of liboptimize
Merge pull request !6351 from lixian/master
|
5 years ago |
mindspore-ci-bot
|
c5c8b9ae19
|
!6405 fix debug build error
Merge pull request !6405 from zhaozhenlong/lite/issue/inline_remove2
|
5 years ago |
zhaozhenlong
|
e08d56fcb1
|
fix -d build error
|
5 years ago |
fuzhiye
|
2901e352b8
|
add determination of malloc (if is nullptr)
|
5 years ago |
lixian
|
1ff44aec02
|
fix function call of liboptimize
|
5 years ago |
mindspore-ci-bot
|
c74c64e0c2
|
!6323 [MS][LITE][CPU]enable int8 kernel on arm32
Merge pull request !6323 from lixian/master
|
5 years ago |
lixian
|
dcaf76a800
|
enable int8 kernel on arm32
|
5 years ago |
fuzhiye
|
85b916ecec
|
modify int8 pooling
|
5 years ago |
mindspore-ci-bot
|
a92e444f21
|
!6268 cmake optimze for mindspore lite
Merge pull request !6268 from zhanghaibo/master
|
5 years ago |
zhanghaibo5@huawei.com
|
56e3736934
|
optimize cmake file for mindspore lite
|
5 years ago |
mindspore-ci-bot
|
e184972cdf
|
!6266 [MSLITE] Support per-channel quantization of int8 matmul on arm32 platform
Merge pull request !6266 from zhanyuan/dev
|
5 years ago |
ling
|
a56ba41c94
|
[MSLITE][Develop] delete useless matmul func
|
5 years ago |
zhanyuan
|
cae7a22b51
|
Support per-channel quantization of int8 matmul for arm32 platform
|
5 years ago |
mindspore-ci-bot
|
966c52b7f6
|
!6194 [MSLITE][Develop]Conv1x1 public preTrasn neon code -> .S
Merge pull request !6194 from ling/conv1x1
|
5 years ago |
ling
|
96d01f17ec
|
[MSLITE][Develop]Conv1x1 preTrasn neon code -> .S
|
5 years ago |
mindspore-ci-bot
|
07422d5438
|
!6202 [MSLITE][Develop] arm cpu int8 conv depthwise support arm32
Merge pull request !6202 from yangruoqi713/int8_arm32
|
5 years ago |