gong chen
a6dfa281ea
Init GraphKernel.
...
- It provides a unified style to express graph and kernel for user.
- It provides a unified IR to represent graph and kernel for developer.
- It breaks the boundary between graph and kernel.
- It provides more opportunities to do compile optimization.
2020-06-20 22:31:54 +08:00
yujianfeng
7ad877a948
Add Split fission pass
2020-06-20 14:28:25 +08:00
huanghui
4acb61d59d
code review fix for buffer fusion
2020-06-15 09:35:16 +08:00
huanghui
118496b3ec
enhance insert memcpy
2020-06-10 18:00:08 +08:00
WilliamLian
9808e47663
change checkAicpu to CheckAICPU & add charge Scalar function to charge the input or output is scalar
2020-06-09 09:31:43 +08:00
huanghui
b4c0ed4b36
add signle batchnorm fission pass
2020-06-08 10:08:23 +08:00
chujinjin
7465abc798
optimize transdata for pynative
2020-06-05 14:10:55 +08:00
mindspore-ci-bot
c51d90d84e
!1767 Move LayerNormGrad split pass ahead of kernel select
...
Merge pull request !1767 from huanghui/LayerNormGrad-split-pass
2020-06-05 09:13:24 +08:00
yujianfeng
e87ac6525e
Add batch norm fusion pattern for mix precision
2020-06-04 20:31:15 +08:00
huanghui
cf87218fb7
place layernormgrad split pass before kernel select
2020-06-04 19:39:40 +08:00
huanghui
d1cec14a0c
add 2 pattern for softmaxgradext fusion pass
2020-06-01 18:48:41 +08:00
mindspore-ci-bot
cc9c004bc1
!1696 Enable ConfusionMulGrad fusion pass
...
Merge pull request !1696 from huanghui/TMP
2020-05-30 14:44:29 +08:00
mindspore-ci-bot
59683a1d90
!1692 Fix topk bug for fasterrcnn
...
Merge pull request !1692 from meixiaowei/master
2020-05-30 09:56:40 +08:00
huanghui
71acaa5300
enable ConfusionMulGrad fusion pass in bert only
2020-05-30 09:46:21 +08:00
mindspore-ci-bot
387bcece9e
!1666 Add 5 patterns for AdamApplyOneWithDecay fusion pass
...
Merge pull request !1666 from huanghui/AdamApplyOneWithDecay-fusion-pass
2020-05-30 09:43:07 +08:00
huanghui
ff05aa1faa
add newly 5 patterns for AdamApplyOneWithDecayRule fusion pass
2020-05-29 19:49:44 +08:00
yujianfeng
e6f1cfa581
Check the input size of BatchNorm before fission in bert
2020-05-29 19:34:58 +08:00
meixiaowei
1778ec0135
fix topk bug
2020-05-29 19:04:18 +08:00
huanghui
99ca6b3e80
add SoftmaxGradExt fusion pass
2020-05-29 14:30:52 +08:00
mindspore-ci-bot
04398cf88e
!1433 add tensor_minnie and separate py from ir
...
Merge pull request !1433 from leopz/tensor_base
2020-05-28 16:00:56 +08:00
leopz
4508134ceb
add tensor_minnie and separate py from ir
2020-05-28 14:26:26 +08:00
yujianfeng
ee087bdf60
Check the size of topk input names before converting input to attr
2020-05-28 10:29:21 +08:00
mindspore-ci-bot
2215e3267f
!1419 remove old buffer fusion pass
...
Merge pull request !1419 from Etone.Chan/Resnet50
2020-05-26 09:16:06 +08:00
etone-chan
42d724d8b4
remove old buffer fusion pass
2020-05-25 21:25:45 +08:00
huanghui
a42dd21430
add 3 pattern for lamb_next_mv_rule fusion pass
2020-05-25 14:18:22 +08:00
huanghui
f16ff539ba
refactor multiple patterns pass
2020-05-23 09:40:21 +08:00
huanghui
eaff850f11
add 3 patterns for lamb_next_mv_with_decay_rule pass
2020-05-21 17:00:15 +08:00
huanghui
5a68eba585
Refactor LambNextMVWithDecayRule fusion pass
2020-05-21 10:08:50 +08:00
huanghui
c4af71e236
add LarsV2 fission pass
2020-05-18 11:30:45 +08:00
huanghui
709828a98b
add BatchNormGrad2BNInferGrad pass
2020-05-15 14:37:50 +08:00
mindspore-ci-bot
42ba885e58
!1098 Add BatchNorm2BNInfer pass
...
Merge pull request !1098 from huanghui/add-batchnorm-bninfer-pass
2020-05-15 14:33:09 +08:00
huanghui
6db71e6576
add batchnorm2bninfer pass
2020-05-15 09:56:38 +08:00
yujianfeng
2b061c84c7
Add batch_norm_grad infer fisson
2020-05-15 09:44:40 +08:00
yujianfeng
6e89ebe6f0
Add BatchNorm fusion pattern with mix precision
2020-05-14 21:28:08 +08:00
etone-chan
c4a5bfb787
add matmul eltwise buffer fusion pass
2020-05-14 10:36:23 +08:00
mindspore-ci-bot
86d4797399
!1079 Convert AiCpu kernel when aicore not supported
...
Merge pull request !1079 from lianliguang/convert-to-AICPU-when-AiCore-not-supported-kernel
2020-05-13 15:15:00 +08:00
WilliamLian
691b0648e3
convert unsupported kernel in aicore to aicpu
2020-05-13 11:52:26 +08:00
yujianfeng
85ff90c237
Add input shape condition for transpose_reshape fusion pass
2020-05-12 22:54:21 +08:00
huanghui
12eaaf710b
mul_add_fusion pass supports when add's 2nd is mul
2020-05-08 14:28:39 +08:00
YuJianfeng
aa6f808616
Add batch norm bert fission pass
2020-05-06 16:33:46 +08:00
changzherui
99c9b48a4d
delete TestHWBatchNormGradSplit
2020-05-04 19:39:17 +08:00
YuJianfeng
7185961e89
Enable BatchNorm fusion pass
2020-04-30 15:29:23 +08:00
huanghui
0de0aee54e
Add BatchNormGrad split pass
2020-04-29 15:28:23 +08:00
kswang
e6f36e050f
change cpu kernel dir
2020-04-27 21:14:03 +08:00
mindspore-ci-bot
e537a70808
!657 Fix confusionmulgrad fusion pass cannot work
...
Merge pull request !657 from huanghui/fix-confusionmulgrad-fusion-pass
2020-04-27 11:10:25 +08:00
mindspore-ci-bot
3d9f1087c6
!695 Check topk supported before converting input to attr
...
Merge pull request !695 from YuJianfeng/master
2020-04-26 21:38:43 +08:00
YuJianfeng
ce2a13fcda
Check topk supported before converting input to attr
2020-04-26 18:40:17 +08:00
mindspore-ci-bot
fb5cfe31d5
!430 add buffer fusion bnupdate eltwise pass
...
Merge pull request !430 from Etone.Chan/bufferfusion
2020-04-26 17:40:20 +08:00
Etone.Chan
d2727d0504
add buffer fusion bnupdate eltwise pass
2020-04-26 17:17:42 +08:00
huanghui
230e77f923
fix confusionmulgrad fusion pass cannot work
2020-04-26 15:35:25 +08:00