chenyemeng
|
5f68aaa0c7
|
rm compile macro
|
4 years ago |
zhengyuanhua
|
2697519926
|
profiling training trace
|
4 years ago |
mindspore-ci-bot
|
4c639c3b4b
|
!794 remove redundant parentheses
From: @xchu42
Reviewed-by: @wqtshg,@ji_chen
Signed-off-by: @ji_chen
|
4 years ago |
mindspore-ci-bot
|
b65e4eb25f
|
!772 GeTensor aligned addr & zero copy support
From: @chen_yemeng
Reviewed-by: @sheng-nan,@wqtshg
Signed-off-by: @wqtshg
|
4 years ago |
chuxing
|
26a50cb9ba
|
remove redundant parentheses
|
4 years ago |
mindspore-ci-bot
|
ca0f3de34d
|
!785 Fixing log text and NoOp callback
From: @xchu42
Reviewed-by: @ji_chen,@wqtshg
Signed-off-by: @ji_chen
|
4 years ago |
chuxing
|
15377e9b97
|
fix typo & NoOp
|
4 years ago |
mindspore-ci-bot
|
acf797db55
|
!763 dynamic shape dump
From: @youui
Reviewed-by:
Signed-off-by:
|
4 years ago |
weiyang
|
6cf3a44a9f
|
dynamic shape
|
4 years ago |
chenyemeng
|
861c0ee1bb
|
GeTensor aligned addr & zero copy support
|
4 years ago |
王涛
|
4207a54092
|
回退 'Pull Request !689 : GeTensor aligned addr & zero copy support'
|
4 years ago |
chuxing
|
509170b150
|
prepare for online infer
|
4 years ago |
chenyemeng
|
a3c8c05e0a
|
Merge remote-tracking branch 'upstream/development' into modify_tensor
|
4 years ago |
lwx897429
|
62bb799d97
|
Add keep_dtype attribute on operators to keep precision unchanged.
|
4 years ago |
chenyemeng
|
87021d3f20
|
GeTensor aligned addr & zero copy support
|
4 years ago |
chenyemeng
|
87224a512a
|
GeTensor aligned addr & zero copy support
|
4 years ago |
zhengyuanhua
|
60fd661c3e
|
profiling data add iter num
|
4 years ago |
dajunli
|
a8cfa75314
|
change hccl inference
|
4 years ago |
chuxing
|
f28de88aaf
|
fix compile error
|
4 years ago |
chuxing
|
bed5633f8b
|
Calc op running param by GE
|
4 years ago |
dongduo
|
84253df880
|
Fix code check
|
4 years ago |
dongduo
|
253f63b791
|
Fix code check
|
4 years ago |
mindspore-ci-bot
|
7462692274
|
!534 for cust aicou
From: @youui
Reviewed-by: @wqtshg,@xchu42,@ji_chen
Signed-off-by: @ji_chen
|
4 years ago |
weiyang
|
8291221f64
|
fix cust aicpu
|
4 years ago |
lichun
|
b8e82bb16e
|
inference supports dynamic shape
|
4 years ago |
wangzhengjun
|
7fe2506953
|
modify for static check 2
|
4 years ago |
dongduo
|
1bdbc03b74
|
Fix code check
|
4 years ago |
mindspore-ci-bot
|
73169daa68
|
!530 Fixing static check
From: @xchu42
Reviewed-by: @ji_chen,@wqtshg
Signed-off-by: @ji_chen
|
4 years ago |
chuxing
|
a31e700c3a
|
fixing static check
|
4 years ago |
mindspore-ci-bot
|
368931a59f
|
!517 static check
From: @selfws
Reviewed-by: @xchu42,@ji_chen
Signed-off-by: @ji_chen
|
4 years ago |
mindspore-ci-bot
|
16ee0ea368
|
!513 decouple cce and resovle new changes support windows
From: @taoxiangdong
Reviewed-by: @youui,@xchu42,@ji_chen
Signed-off-by:
|
4 years ago |
wangzhengjun
|
fde6b4c653
|
modify for static check
|
4 years ago |
王涛
|
87704d8512
|
回退 'Pull Request !499 : inference supports dynamic shape'
|
4 years ago |
taoxiangdong
|
7d8aa2e94a
|
graphengine support windows new changes
|
4 years ago |
taoxiangdong
|
55fd4c4c14
|
decouple cce in graphengine
|
4 years ago |
lichun
|
209767e0ca
|
inference supports dynamic shape
|
4 years ago |
cclworkaccount
|
35f7bb4c06
|
!475 GE添加reduce算子定义
* update ge/graph/manager/util/hcom_util.cc.
* !1 update ge/graph/load/new_model_manager/davinci_model.cc.
* update ge/graph/build/stream_allocator.cc.
* update ge/graph/load/new_model_manager/davinci_model.cc.
* update ge/graph/manager/util/hcom_util.cc.
* update inc/framework/common/types.h.
* update ge/hybrid/node_executor/hccl/hccl_node_executor.cc.
* update ge/graph/optimize/mem_rw_conflict_optimize.cc.
* update ge/graph/manager/util/hcom_util.cc.
* update ge/graph/load/new_model_manager/task_info/hccl_task_info.cc.
* update ge/graph/build/stream_allocator.cc.
* update ge/graph/build/stream_allocator.cc.
* update ge/common/types.cc.
|
4 years ago |
chuxing
|
6186ac0a22
|
security check & fixing optional inputs
|
4 years ago |
lichun
|
9801866dfc
|
Fix the error of unexpected executor type
|
4 years ago |
unknown
|
71788dcc50
|
No session in tf_singleop_task and add cpu kernels cache.
|
4 years ago |
lichun
|
cb397142ac
|
Feature:infer support dynamic shape
|
4 years ago |
lichun
|
c3f4b95ffd
|
Decoupling of aicore node executor and aicore task compiler
|
4 years ago |
mindspore-ci-bot
|
f01cfe5167
|
!278 Support StatelessIf and StatelessWhile
From: @xchu42
Reviewed-by: @ji_chen,@wqtshg
Signed-off-by: @ji_chen
|
4 years ago |
chuxing
|
54fab4c9aa
|
supporting stateless if/while
|
4 years ago |
mindspore-ci-bot
|
c0e8eee238
|
!259 Skipping aicore operators whose output tensors are all empty
From: @xchu42
Reviewed-by: @sheng-nan
Signed-off-by:
|
4 years ago |
isaactalx
|
a9412057df
|
subgraph netoutput memory bugfix
|
4 years ago |
chuxing
|
ff6f8d8e46
|
skip aicore operators whose output tensors are all empty
|
4 years ago |
weiyang
|
1c10ac4821
|
for cust aicpu
|
4 years ago |
王涛
|
a474fd8810
|
!109 Add properties to determine whether atomic_clean is needed at runtime.
Merge pull request !109 from 赵之轩/development
|
4 years ago |
unknown
|
99e8cc1612
|
support for atomic clear task.
|
4 years ago |