Nigel
/
mindspore
forked from mindspore-Ecosystem/mindspore
0
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
51,586 Commits 24 Branches 0 Tags 3.9 GiB
Commit Graph

608 Commits

Author SHA1 Message Date
yangzhenzhang 9489bc9ce6 modify stridedslice parallel op 2022-08-27 15:10:52 +08:00
hezhenhao1 bc12b4ef87 Add function/tensor api for PReLU, Square, SoftplusGrad operators. 2022-08-25 14:28:31 +08:00
i-robot d222d7393c
!40455 [AutoParallel]optimizer Pipeline parallel 
Merge pull request !40455 from lichen/opt_pipeline_parallel
 2022-08-19 01:15:31 +00:00
lichen 23a3d65995 optimizer_pipeline_parallel 2022-08-17 17:40:26 +08:00
yao_yf 854b7839e5 auto parallel dataset handle in feed mode 2022-08-15 14:41:16 +08:00
yangzhenzhang b83c99ce71 add flag for broadcast net 2022-08-04 17:07:49 +08:00
yangzhenzhang 5fa63f8e46 delete auto parallel flag 2022-07-30 15:19:56 +08:00
Zhang Qinghua f90dcc963d Support isolated side-effect nodes in constant returned func graph. 2022-07-27 10:12:24 +08:00
wangshengnan12@huawei.com 937e65e81a clean_code_auto_parallel 2022-07-21 15:06:05 +08:00
zhangyinxia 3e1320b246 add lamb and clipbynorm pass for ge 2022-07-20 10:54:06 +08:00
yangzhenzhang 25009f9051 modify gather op 2022-07-07 14:59:50 +08:00
i-robot 58c901923a
!36767 modify the generated strategy for conv2d 
Merge pull request !36767 from yangzhenzhang/modify-conv2d-gen-strategy-func
 2022-06-29 01:39:29 +00:00
wzw d309dfd41e add transformer testcase 2022-06-28 20:01:29 +08:00
yangzhenzhang 9b50d2cfee modify conv2d gen strategy func 2022-06-28 16:32:15 +08:00
wangshengnan12@huawei.com bdf2004ca4 fix_pipeline_compile_error 2022-06-28 09:55:31 +08:00
huangxinjing 055ab431db [AutoParallel]Add virtual mul 
Fix OneHot Error
Add onehot test cast
 2022-06-25 09:06:40 +08:00
huangxinjing d0f2bf37c6 Add cross entory loss 
[AutoParallel]Add sub graph support for parallel training
Fix forward mark error and virtual dataset error
Add Test cases
Add virtual dataset inserts, forbiden the subgraphs as candidates
 2022-06-20 10:53:21 +08:00
i-robot e902c7c731
!35839 fix bug of moe 
Merge pull request !35839 from bichaoyang/master
 2022-06-17 08:15:25 +00:00
yangzhenzhang b63e70f6a9 gen data parallel strategy for conv2d auto parallel 2022-06-16 16:44:30 +08:00
bichaoyang 68398b054a fix bug of moe 2022-06-14 20:04:27 +08:00
yangzhenzhang e5b62b2714 modify gen strategy for matmul op 2022-06-09 19:50:31 +08:00
i-robot 6f18373b10
!35252 add python implement and st for random gamma 
Merge pull request !35252 from zhangqi/0522_2
 2022-06-02 09:26:10 +00:00
zhangqi d0afaf60f0 add python implement and st for random gamma 2022-06-01 11:20:36 +08:00
yangzhenzhang 3df7174d25 modify gen strategy for parallel op 2022-05-31 18:41:42 +08:00
i-robot 291b5faa98
!33925 parallel_initializer_seed 
Merge pull request !33925 from yao_yf/auto_parallel_initializer_seed
 2022-05-25 02:03:16 +00:00
i-robot 4f0b1dc60e
!34511 Add CPU support for KLDivLoss op 
Merge pull request !34511 from zhuyuxiao/I51VMV
 2022-05-23 11:06:37 +00:00
yao_yf 91dabdce17 parallel initializer seed set 2022-05-23 14:24:29 +08:00
zhuyuxiao ce16351d16 add CPU support for KLDivLoss op 2022-05-20 15:34:31 +08:00
i-robot 71f2ee7172
!34555 [AutoParallel]Parallel_adapt_simple_control_flow 
Merge pull request !34555 from lichen/adapt_control_flow
 2022-05-20 01:31:48 +00:00
i-robot 527539ef7c
!34409 unsorted segment prod gpu inference 
Merge pull request !34409 from ling/core
 2022-05-20 01:20:54 +00:00
i-robot 67033817ff
!34295 Optimiz lamb ascend kernel 
Merge pull request !34295 from wanghenchang/lamb-optimiz-ascend
 2022-05-20 01:20:25 +00:00
wangchangheng 6c879b5554 optimize lamb 2022-05-19 19:18:16 +08:00
ling 437deb9da2 unsorted segment prod gpu inference 2022-05-19 18:59:19 +08:00
Zhang Qinghua d7762f1c8d Add bprop_return_sparse flag for sparse bprop primitive and remove context option: 'enable_sparse'. 2022-05-19 17:31:57 +08:00
wangshengnan12@huawei.com 48065b72a1 adapt_control_flow 2022-05-19 15:50:29 +08:00
bichaoyang 98d4b75d1b reminder when missing 'pipeline_end' 2022-05-16 11:11:00 +08:00
i-robot 6c78153ef8
!33948 add softmax mul parallel tests 
Merge pull request !33948 from wanhanyang/new_softmax
 2022-05-06 09:00:41 +00:00
i-robot 3409f924bc
!33947 Add softmax gather parallel tests 
Merge pull request !33947 from wanhanyang/test_softmax_gather
 2022-05-06 08:31:38 +00:00
i-robot 8f565f098e
!33904 use _inner_ops shard instand of array_ops range 
Merge pull request !33904 from yanzhenxiang2020/range_shard_inner
 2022-05-06 08:10:46 +00:00
Wan Hanyang 2cf05811b8 add test softmax gather net tests 2022-05-06 14:59:27 +08:00
Wan Hanyang d16b2c15b1 Add mul softmax net tests 2022-05-06 14:53:30 +08:00
i-robot 1b01960935
!33923 回退 'Pull Request !33791 : Add gather with softmax parallel tests' 
Merge pull request !33923 from suteng/revert-merge-33791-master
 2022-05-06 05:42:18 +00:00
suteng 5c9f240735
回退 'Pull Request !33791 : Add gather with softmax parallel tests' 2022-05-06 03:20:35 +00:00
suteng 9c5d75d416
回退 'Pull Request !33790 : Add parallel fused softmax parallel tests' 2022-05-06 03:20:18 +00:00
yanzhenxiang2020 04c5a582da use _inner_ops shard instand of array_ops range 2022-05-06 10:36:57 +08:00
i-robot 43fd864c10
!31745 Enable if parallel call flag by default 
Merge pull request !31745 from xychow/enable-if-parallel-call-by-default
 2022-05-05 09:28:39 +00:00
i-robot 3d1060dcdb
!33790 Add parallel fused softmax parallel tests 
Merge pull request !33790 from wanhanyang/softmax2
 2022-04-29 13:02:50 +00:00
Su Teng 1686559b0d add gather with softmax parallel tests 2022-04-29 18:18:01 +08:00
Su Teng 1e1eaedec7 add parallel fused softmax tests 2022-04-29 18:14:43 +08:00
zhousiyi d4aba77d3d enable if parallel call transformation by default; fix test case ut/python/pipeline/parse/test_fix_bug.py::test_fix_expanddims_loss_scale 
Check only if all branch of middle graph's output are scalar, parallel_if transformation should be ignored

fix test_auto_monad.py::test_while_if geswitch input shape check failed

fix ut/python/pipeline/parse/test_super.py::test_super_cell core dump as varg and kwarg parameter are removed but those flags are not set, it will cause FuncGraph::GetPositionArgsCount be negative

skip test_control_flow in ut/python/paralle as Parallel team is working on contro flow adaption.

fix ut/python/ir/test_row_tensor.py::test_row_tensor_with_control_flow_if

fix for kwonlyargs when eliminate_parameter and add testdases

fix cannot join for st/control/inner/test_012_if_in_for.py

fix pclint errors

fix test/st/scipy_st/sparse/test_linalg.py, if the output cnode of former funcgraph is Depend CNode with input 1 as middle funcgraph call, then keep the output of former funcgraph as Depend other than middle funcgraph tail call.

log as exception

If the output of middle funcgraph contain PyInterpret CNode, then ignore
the transformation as the Value in Abstract will be kAnyValue after join
of the PyInterpret CNode value.

Backend cannot support for RealDiv with Int64, so change one operand to Float
 2022-04-29 08:22:56 +00:00