7347157+joylvliang@user.noreply.gitee.com
39e70e8519
fix_bug_of_tensor_copy_canot_work_in_int64_with_D
4 years ago
wanyiming
a124ec4de7
add dynamic_decay
4 years ago
zhoufeng
8e97ce5cd2
build core.so
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
4 years ago
Margaret_wangrui
7d1f3d71ba
Optimize error message
4 years ago
zhangzhaoju
c5b2996b3f
Resuse ConvertData of GenerateArgumentsKey
4 years ago
zhangzhaoju
722086435d
1. fix:unexpected eiminate of tile
2. fix:delay memory allocat in NewTensorFilledWithData
4 years ago
yujianfeng
7c808ee792
Add some bprop mindir files
4 years ago
yujianfeng
f2bd7de442
Add automatic detection for the changes of python scripts when enable compile cache
4 years ago
l00591931
9353bd10a4
Enable tuple as variable in optimizer
4 years ago
yujianfeng
fc769a79fe
Fix the unloadable bprop mindir files
4 years ago
yujianfeng
195563c67f
Fix the random failure of ut tests in ci
4 years ago
yujianfeng
42079ac087
Change dir hash to file hash in bprop mindir
4 years ago
wanyiming
8b40e77201
ut_fixed
4 years ago
i-robot
bbdacd41f4
!20585 add security isolate for save_graphs
Merge pull request !20585 from huanghui/add-security-isolate-for-DumpIR
4 years ago
huanghui
ba66c0d491
add security isolate for save_graphs
4 years ago
dinglinhe
9a8c1b3c73
array_ops error recti and others
4 years ago
yujianfeng
752bdde06d
Optimize of exception log of AD
4 years ago
yujianfeng
dfe1b05b28
Get bprop from python when the bprop mindir dir does not exist
4 years ago
yao_yf
39055af6e4
recompute interface modify
4 years ago
Zhang Qinghua
a137fa1d0b
Optimize the Executors routines.
- Fix the key generating.
- Distinguish the executors.
4 years ago
yujianfeng
712b9bd013
convert some ops bprops to mindir
4 years ago
ms_yan
36a8886ca2
Revert "[feat] [assistant] [I3T96T] add new Dataset operator CMUARCTICDataset"
This reverts commit b077aa1cab .
Revert "[feat] [assistant] [I3T96X] add new Dataset operator LibriSpeechDataset"
This reverts commit 4e6f7dc97d .
delete pass_registry_test.cc
comment hiai_nlu_model_multi.pb related line
4 years ago
djc
b077aa1cab
[feat] [assistant] [I3T96T] add new Dataset operator CMUARCTICDataset
4 years ago
djc
4e6f7dc97d
[feat] [assistant] [I3T96X] add new Dataset operator LibriSpeechDataset
4 years ago
yao_yf
5277b229be
add cell comm recompute interface
4 years ago
yujianfeng
cf4121126a
Only replace the primal user with tuple_getitem once
4 years ago
i-robot
ed4c9682b5
!20505 limit the scope of lift free variable
Merge pull request !20505 from xychow/limit-lift-scope
4 years ago
zhousiyi
597f29ea7d
limit the lambda lift scope to the func_graph passed
4 years ago
zhousiyi
b590f6d929
don't replace u with fprop_u in primal_graph and bprop_fg
4 years ago
zhousiyi
a5b4e7bbf8
lift fv before grad except weight, then convert
switch(cond, partial(g1, xs), partial(g2, ys))(Zs)
to
switch(cond, g1, g2)(Xs, Ys, Zs)
switch_layer(index, make_tuple(partial(g1, xs), partial(g2, ys)))(Zs)
to
switch_layer(index, make_tuple(g1, g2))(Xs, Ys, Zs)
put Zs at last when unifyparameter as it may have u-monad or io-monad
use joined args other than broadened one as some extra parameter which is not a parameter of while_header can be add to while_body
inline fprop_switch forcely
reorder the parameter if one of the parameter is Monad when incorporate call
incorporate switch tuple_getitem if item 0 of tuple is EnvInstance or
item 1 of tuple is bprop function
addn with shape() and shape(1)
remove context_ from FuncGraphEvaluator to make it re-entry able to resolve evaluator stuck issue because of re-entry of the same FuncGraphEvaluator
4 years ago
l00591931
8ae5d7cc84
Fix compile cache for resnet50
4 years ago
huanghui
53e32077c1
fix code check
4 years ago
zhousiyi
74d500a756
If a fv cnode is not mapped but belong to primal_graph_, then don't propagate this fv to caller, it should mapped instead
5 years ago
buxue
2551f5ea7e
support name or attribute ast.Expr
5 years ago
l00591931
680324f225
Change make tuple in core.ops
5 years ago
yujianfeng
70cc548e32
Fix calling recompute api after compiling
5 years ago
mindspore-ci-bot
b189f177bb
Change tuple_getitem to TupleGetItem and some other ops, merge from r1.1 to master
5 years ago
l00591931
9ec100d069
Change TensorAdd to Add, from r1.1 to master
5 years ago
yujianfeng
728fac6c9f
Expand J for innermost graph first when the graph also contains J primitive
5 years ago
yujianfeng
e1aa34a030
Ignore the graphs with defer_inline flag when specializing on graph arguments
5 years ago
mindspore-ci-bot
37e3b6082f
!6002 Reduce inline passes traversal by skipping inline action if no pre-ad python pass exists
Merge pull request !6002 from BowenK/pre_ad
5 years ago
Jiaqi
94d63b90f4
remove sens parameter
5 years ago
BowenK
e482e4e8bf
reduce pass traversals if no pre-ad python pass exists
5 years ago
BowenK
1bdb26f9e8
Warming up python pass by adding inline passes before it
5 years ago
BowenK
d6fb7d2db1
Remove debug drawing and printing to boost compile performance; re-opt after python pass to boost training; fix NewParameter tensor clone
5 years ago
BowenK
641d12d6d9
python pass pattern renaming and interface tweaking
5 years ago
mindspore-ci-bot
8d693306f4
!4126 Add new parameter
Merge pull request !4126 from BowenK/new_parameter
5 years ago
BowenK
e7c6b7e66a
Add NewParameter and Imm patterns
5 years ago
李嘉琪
c65ea1687b
modify error type
5 years ago
Wei Luning
d4d6457ea7
init parameter data by defaultOnly keep no data as MetaTensor in auto parallel mode
5 years ago