lizhenyu
|
e33e2fc1a0
|
add bn fusion pass
|
5 years ago |
mindspore-ci-bot
|
58523a41fe
|
!4392 use builtin float16 for arm
Merge pull request !4392 from xychow/use-float16-in-arm-neon
|
5 years ago |
zhoufeng
|
663278112f
|
optimize code compile performance
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
|
5 years ago |
huanghui
|
71d839dec7
|
fix do fission when UnsegmentSum's input0 shape size is 1
|
5 years ago |
mindspore-ci-bot
|
d341461baa
|
!4510 GPU fix codex for replace_bn_grad_cast
Merge pull request !4510 from VectorSL/fix-codex
|
5 years ago |
zhousiyi
|
e1aa49a4b7
|
use built-in float16 in arm_neon.h for lite arm
|
5 years ago |
WilliamLian
|
6760d9976d
|
add reshape type to tensor
|
5 years ago |
VectorSL
|
9dd0282a17
|
git update relplace_bn_grad_cast
|
5 years ago |
mindspore-ci-bot
|
61a2274787
|
!4359 fix layernorm grad error in pynative mode
Merge pull request !4359 from chujinjin/fix_layernorm_grad_error_in_pynative
|
5 years ago |
mindspore-ci-bot
|
9e5f4182a8
|
!4329 Revert the AdamApplyOneAssign fusion pass before op registration
Merge pull request !4329 from YuJianfeng/adam_assign
|
5 years ago |
chujinjin
|
9cc9f6b470
|
fix layernorm grad error in pynative
|
5 years ago |
huanghui
|
b8d7f6d77f
|
add UnsortedSegmentSum fission pass
|
5 years ago |
yujianfeng
|
cb03bbba44
|
Disable the AdamApplyOneAssign fusion pass before op registration
|
5 years ago |
mindspore-ci-bot
|
196fdbe16e
|
!4280 adapt input to attr between cpu and aicpu embeddinglookup
Merge pull request !4280 from wuxuejian/embedding_input_adapt
|
5 years ago |
mindspore-ci-bot
|
60dd4d7185
|
!4253 fix bug of merge cast to next op
Merge pull request !4253 from lianliguang/master
|
5 years ago |
wuxuejian
|
5e6b1d42da
|
adapt input to attr between cpu and aicpu embeddinglookup
|
5 years ago |
mindspore-ci-bot
|
0115876363
|
!4197 Add AdamApplyOneAssign fusion pass
Merge pull request !4197 from YuJianfeng/adam_assign
|
5 years ago |
WilliamLian
|
8a19a7c3de
|
fix bug of merge cast to next op
|
5 years ago |
mindspore-ci-bot
|
a0bfeedfa5
|
!4232 modify the condition of pattern match in bnupdate + eltwise fusion pass
Merge pull request !4232 from Etone.Chan/August
|
5 years ago |
mindspore-ci-bot
|
dc961e461e
|
!4225 Fix bug avoid multi attr value be eliminated in pynative mode
Merge pull request !4225 from JoyLvliang/fix-bug-avoid-multi-attr-value-be-eliminated-in-pynative-mode
|
5 years ago |
etone-chan
|
18c83637f1
|
modify the condition of pattern match in bnupdate + eltwise fusion pass
|
5 years ago |
mindspore-ci-bot
|
0a1fac9246
|
!3952 add matmul confusiontranspose buffer fusion
Merge pull request !3952 from Etone.Chan/August
|
5 years ago |
lvliang
|
e1a3c39fac
|
fix-bug-avoid-multi-attr-value-be-eliminated-in-pynative-mode
|
5 years ago |
etone-chan
|
bae2ada744
|
add matmul confusiontranspose buffer fusion
|
5 years ago |
yujianfeng
|
8a77751988
|
Add AdamApplyOneAssign and AdamApplyOneWithDecayAssign fusion pass
|
5 years ago |
jjfeing
|
b86b5a836a
|
add split pass on pynative mode
|
5 years ago |
huanghui
|
d7f28a72df
|
fix when ReduceMin's axis set with one int number
|
5 years ago |
mindspore-ci-bot
|
607cb58ae5
|
!3796 Gpu AdamWeightDecay fusion
Merge pull request !3796 from chenweifeng/AdamWeighDecayFusionFix
|
5 years ago |
mindspore-ci-bot
|
172a468704
|
!3968 Ignore node with _side_effect attr in cse
Merge pull request !3968 from YuJianfeng/cse
|
5 years ago |
mindspore-ci-bot
|
b045f47428
|
!3983 Add ReduceMin fission pass
Merge pull request !3983 from huanghui/reduce-min-fission-pass
|
5 years ago |
mindspore-ci-bot
|
52689a7dcf
|
!3938 decoupling core and context
Merge pull request !3938 from liubuyu/master
|
5 years ago |
zhoufeng
|
ca7154a548
|
graph compile performance optimization
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
|
5 years ago |
huanghui
|
30000fdb52
|
add ReduceMin fission pass
|
5 years ago |
liubuyu
|
d81862a916
|
decoupling core and context
|
5 years ago |
yujianfeng
|
bff3600fcf
|
ignore node which has _side_effect attr in cse
|
5 years ago |
yujianfeng
|
57cb1eeb14
|
Performance optimization of embedding_lookup
|
5 years ago |
mindspore-ci-bot
|
953f91cef8
|
!3739 not eliminate memcpy when nexe node is graph output
Merge pull request !3739 from laiyongqiang/eliminate_memcpy_bug
|
5 years ago |
mindspore-ci-bot
|
c9453a614e
|
!3239 Insert concat for outputs of AllGather
Merge pull request !3239 from YuJianfeng/allgther
|
5 years ago |
liubuyu
|
a499d4e47c
|
decoupling core and debug
|
5 years ago |
wilfChen
|
fbc696ad75
|
gpu AdamWeightDecay
|
5 years ago |
laiyongqiang
|
2458431750
|
not eliminate memcpy when nexe node is graph output
|
5 years ago |
mindspore-ci-bot
|
0df4b11487
|
!3709 GPU update bng pass
Merge pull request !3709 from VectorSL/update-bng-pass
|
5 years ago |
yujianfeng
|
47ab812edb
|
Insert concat for AllGather outputs
|
5 years ago |
mindspore-ci-bot
|
f87983833d
|
!3636 Fix opt unnecessary catch
Merge pull request !3636 from Kang/opt
|
5 years ago |
wuyongkang
|
f1e639a86d
|
Fix opt unnecessary catch
|
5 years ago |
VectorSL
|
9b0602747c
|
gpu update bng pass
|
5 years ago |
limingqi107
|
af39ca8252
|
modify the wrong word
|
5 years ago |
mindspore-ci-bot
|
1b69923472
|
!3643 Throw exception if different communication ops which are divided to the same segement share the same input
Merge pull request !3643 from huanghui/communication-op-fusion
|
5 years ago |
huanghui
|
311d8ea1f9
|
add exception when different communication op in one segment shared the same input
|
5 years ago |
laiyongqiang
|
d99786e938
|
fix refnode input type assign
|
5 years ago |