vilc
/
mindspore2022
forked from huawei/mindspore2022
0
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
42,205 Commits 18 Branches 22 Tags 3.6 GiB
Commit Graph

18639 Commits

Author SHA1 Message Date
mxm d375724055 fixed: PrimitiveToInferImplMap map is global, and key of the map PrimitivePtr also a global variable. If key is initialized later than the map initialized during compilation, will cause the primitive map initialize failed. Variable initialization order is not guaranteed during compilation. 2020-04-10 19:10:45 +08:00
Alexey Shevlyakov 6d1ea7af8e remove make_unique.h 2020-04-10 19:10:45 +08:00
Zirui Wu 71b81c8f1b implemented multi-thread index writer for mindrecord 
num threads cannot be more than num shards

minor fix

clang style fix

address review comments
 2020-04-10 19:10:45 +08:00
jonyguo 6690a7fd7a fix: error info is not exactly when column list invalid 2020-04-10 19:09:54 +08:00
anzhengqi 82896411e0 modify part of comments 2020-04-10 19:09:54 +08:00
jinyaohui e01df47915 modify set_dataset_mode_config api param 2020-04-10 19:09:54 +08:00
liuxiao 599543932c Add pack and unpack 2020-04-10 19:09:21 +08:00
jojobugfree 89f0b3b1bb profiling feature enhancement 2020-04-10 19:08:25 +08:00
Xiaoda Zhang 7798c85e70 This commit is to separate the computation cost and memory cost in auto_parallel. Some related memory correction is removed. 2020-04-10 19:08:25 +08:00
Jonathan Yan f01098bc12 remove ENABLE_MINDRECORD flag 2020-04-10 19:08:25 +08:00
jonyguo 6dc6d6bc83 fix: when use MindDataset block_reade=True hung 2020-04-10 19:08:25 +08:00
dengwentao c6c8c94d3f updata mkl-dnn link and md5 2020-04-10 19:08:25 +08:00
wanghua a8f95e6d2c modify bert test file 2020-04-10 19:05:22 +08:00
wanghua 3aa51f35c1 fix bert precison bug 2020-04-10 19:05:22 +08:00
kingfo 93fe493cf4 fix ME+GE compile error 2020-04-10 19:05:22 +08:00
zhangz0911gm 23c21e191f Add FloorMod, Acosh in ME 2020-04-10 18:58:39 +08:00
Wei Luning 2ba026dbf2 remove ge depend in cpu 2020-04-10 18:57:16 +08:00
zhaoting 9b5d4eff89 add RMSProp optimizer 2020-04-10 18:51:41 +08:00
lichenever 07449cd1cc fix two cast bug in auto parallel 2020-04-10 18:44:53 +08:00
zhaozhenlong 1984e4a1ff add operator diag and diag_part 2020-04-10 18:37:03 +08:00
leonwanghui 4c2aa41f1d 回退 'Pull Request !17 : [AutoParallel]Fix bug in the case of two cast' 2020-04-10 18:34:27 +08:00
lichenever f0f55ad7e8 fix_cast_bug 2020-04-10 18:33:53 +08:00
lizhenyu 8cd3308a4c change package name akg to _akg 2020-04-10 17:16:45 +08:00
义峰潘 2bef22d8a3 fix bprop cache caused error with variable params 2020-04-10 12:21:40 +08:00
chenfei d1a0ded6c2 use first depend create parameter 2020-04-10 11:06:24 +08:00
mindspore-ci-bot 268d358a1d !187 refactor OperatorCostPtr in OperatorInfo 
Merge pull request !187 from chentingting/refactor_OperatorCostPtr_in_OperatorInfo
 2020-04-10 09:21:25 +08:00
mindspore-ci-bot 1b3b3b1a1c !198 [opt] momentum duplicate mul constant 
Merge pull request !198 from biffex/momentum-duplicate-mul-constant
 2020-04-10 08:37:41 +08:00
Zirui Wu 19f2ea6d41 change int32 to int64 to avoid overflow in batch 
fix clang

pybind fix
 2020-04-09 11:23:43 -04:00
biffex 62bbf560c6 constant duplicate mul for momentum 2020-04-09 22:28:00 +08:00
c00425699 b413638f23 refactor OperatorCostPtr in OperatorInfo 2020-04-09 20:37:52 +08:00
mindspore-ci-bot 50cb2ad944 !192 Modify api name Stack -> Pack, Unstack -> Unpack in ME. 
Merge pull request !192 from liuxiao/temp
 2020-04-09 20:08:42 +08:00
mindspore-ci-bot 002fd9a446 !202 add variable ASCEND_CUSTOM_PATH for customized Ascend installation 
Merge pull request !202 from yanghaoran/master
 2020-04-09 19:54:46 +08:00
yanghaoran 576f65ad48 add custom environment variable ASCEND_CUSTOM_PATH for customized Ascend software installation 2020-04-09 19:07:01 +08:00
liuxiao 0f0f83e292 modified api name Stack -> Pack, Unstack -> Unpack 2020-04-09 19:02:59 +08:00
mindspore-ci-bot cc53ddaeca !171 Add aicpu op register 
Merge pull request !171 from zjun/master
 2020-04-09 18:55:19 +08:00
zjun 16296da5c7 add aicpu opinfo register 2020-04-09 16:33:17 +08:00
kswang bef62db128 add ascend mem pool 2020-04-09 16:04:59 +08:00
mindspore-ci-bot 3bde6404de !27 Rebuild graph before RunGraph if needed 
Merge pull request !27 from Simson/master
 2020-04-09 14:33:29 +08:00
mindspore-ci-bot 0565e4641e !185 modify comment about normal mode 
Merge pull request !185 from jinyaohui/master
 2020-04-09 14:24:28 +08:00
mindspore-ci-bot 2e6e94b2b6 !177 prelu operator support parallel on the channel 
Merge pull request !177 from yao_yf/fix_auto_parallel_prelu
 2020-04-09 14:08:41 +08:00
mindspore-ci-bot 31efc8b088 !172 add mem manager 
Merge pull request !172 from kisnwang/add-resource-manager
 2020-04-09 11:36:33 +08:00
mindspore-ci-bot 086edc133f !65 fix bug of global variable initialization order is not guaranteed during compilation 
Merge pull request !65 from mxm/for_prim_map_init
 2020-04-09 10:19:50 +08:00
zhousiyi 99f22c5b85 fix children issue: missing users of the children's direct users. 2020-04-09 02:18:55 +00:00
kswang fb343bd607 add mem manager 2020-04-09 10:05:45 +08:00
buxue 1d3bb0b731 Develop op MaxPoolWithArgMax 2020-04-09 09:26:56 +08:00
jinyaohui 0790ef8a85 modify comment 2020-04-09 09:23:39 +08:00
simson 88d173cc48 rebuild graph if needed 2020-04-09 09:20:08 +08:00
mindspore-ci-bot cc75cb357c !168 remove mindspore::make_unique and make_unique.h 
Merge pull request !168 from Alexey_Shevlyakov/remove_mindspore_make_unique
 2020-04-09 05:32:52 +08:00
mindspore-ci-bot bd7ac259af !147 Parallelize mindrecord index writer via std::thread 
Merge pull request !147 from ZiruiWu/master
 2020-04-09 05:31:44 +08:00
mindspore-ci-bot 3e36982314 !99 Develop op MaxPoolWithArgMax 
Merge pull request !99 from zhangbuxue/Dock_MaxPoolWithArgMax
 2020-04-08 22:40:00 +08:00
Zirui Wu 5637f80692 implemented multi-thread index writer for mindrecord 
num threads cannot be more than num shards

minor fix

clang style fix

address review comments
 2020-04-08 10:03:19 -04:00
mindspore-ci-bot 68f804255f !174 enhance: the error info is not detail when the column list is invalid by MindDataset 
Merge pull request !174 from guozhijian/fix_false_error_info
 2020-04-08 21:52:00 +08:00
mindspore-ci-bot b35046f559 !173 modify image analysis capabilities comments 
Merge pull request !173 from anzhengqi/modify-obsolete-comments
 2020-04-08 21:38:01 +08:00
mindspore-ci-bot 40d4a4baa3 !20 modify _set_dataset_mode_config api param 
Merge pull request !20 from jinyaohui/master
 2020-04-08 20:58:26 +08:00
mindspore-ci-bot fd7d75aea3 !143 Adapting ops Stack and Unsatck in ME 
Merge pull request !143 from liuxiao/temp
 2020-04-08 20:56:33 +08:00
buxue 7541d3b067 Develop op MaxPoolWithArgMax 2020-04-08 20:47:49 +08:00
yao_yf b5e3fa9593 fix auto parallel prelu 2020-04-08 20:45:08 +08:00
chang zherui 9f427befd1 utf 2020-04-08 20:02:15 +08:00
kingfo cdbef85e5e refactor callback for ge backend 2020-04-08 20:00:06 +08:00
leonwanghui e8d2a6351e Fix some typo errors in mindspore ir module 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-04-08 20:00:06 +08:00
chenhaozhe 36899e0b1d change log level in pipeline_ge 2020-04-08 20:00:06 +08:00
Wei Luning 826a7393db fix InsertGradientOf with class method 2020-04-08 20:00:06 +08:00
panyifeng e8ae576d10 fix grad missing due to indirect dependent free morphism 2020-04-08 20:00:06 +08:00
leonwanghui 48d80d591a Fix some typo errors in pipeline module 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-04-08 20:00:06 +08:00
wanghua d42954652c modify bert test file 2020-04-08 20:00:06 +08:00
guohongzilong 434abe3f8c import comment and function of op print 2020-04-08 20:00:06 +08:00
wanghua 1a82ff0a72 fix bert precison bug 2020-04-08 20:00:06 +08:00
y00445136 26e6138247 add support for bool 2020-04-08 20:00:06 +08:00
zjun 51ca769d09 add new mode for operator info register 2020-04-08 20:00:06 +08:00
yangzhenzhang 8f74aef65d add parallel ops for expand dims 2020-04-08 20:00:06 +08:00
kingfo 7329a1ef87 fix ME+GE compile error 2020-04-08 20:00:06 +08:00
zhaoting a62a9f4395 add YOLOv3 infer scipt and change dataset to MindRecord 2020-04-08 20:00:06 +08:00
guohongzilong db405872e7 pass str to ge.exec.EnableDump 2020-04-08 20:00:06 +08:00
jjfeing 5583532ec6 tr5 run package update 2020-04-08 20:00:06 +08:00
leonwanghui e404187970 Fix some typo errors in session and device module 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-04-08 20:00:06 +08:00
YuJianfeng 024706f951 Optimize depend edge with make tuple input 2020-04-08 20:00:06 +08:00
c00425699 b3fd38c23a fix bug for allreduce fusion and add resnet unit test 2020-04-08 20:00:06 +08:00
lichenever 10312620d2 register not equal distributed op 2020-04-08 20:00:06 +08:00
zhangz0911gm 65cd9d48fd Add FloorMod, Acosh in ME 2020-04-08 20:00:06 +08:00
geekun fb92b5b16e fix custom op bug and add custom op check 2020-04-08 20:00:06 +08:00
Wei Luning b18f634912 remove ge depend in cpu 2020-04-08 20:00:06 +08:00
lichenever 89945c15f5 fix refkey bug for auto parallel 2020-04-08 20:00:06 +08:00
Su Teng 91283d0c74 sort include file in parallel dir 2020-04-08 20:00:06 +08:00
laiyongqiang 3e68560023 fix bug when dump op with trans_flag true 2020-04-08 19:58:14 +08:00
huangdongrun 84e4942ca9 support grad on graph with variable arguments 
use unpack graph primitive instead

add testcases for all grad interface

remove debug log

format code

remove dumpfuncgraph

resolve clang-format

resolve reviews

resolve cpplint

fix review
 2020-04-08 19:58:14 +08:00
zhaoting dcd1f0a504 add RMSProp optimizer 2020-04-08 19:58:14 +08:00
Xiaoda Zhang 9a717aa1f7 change_star_elimination: make the non-identity triangle_eliminatin exact 2020-04-08 19:58:14 +08:00
kswang 0c1aa0d566 add getptr for memreuse 2020-04-08 19:58:14 +08:00
zhaozhenlong 2498d2c749 add atan2 operator 2020-04-08 19:58:14 +08:00
chenhaozhe 99173fb0f5 modify log level in DfGraphManager 2020-04-08 19:58:13 +08:00
Peilin Wang 3c1d35295c TFReaderOp fix, threads will exit after reading necessary amount of rows 
changes from yanpanhui 524009: added set_dataset_size and changed get_dataest_size according to ME requirements

CI fixes
 2020-04-08 19:58:13 +08:00
yangzhenzhang c44e12713b add parallel ops for neg and batchmatmul 2020-04-08 19:58:13 +08:00
Xiaoda Zhang fb00dae17a change star elimination: remove some redundant and checking works 2020-04-08 19:58:13 +08:00
jojobugfree 2c751b75b5 remove cce dependency 2020-04-08 19:58:13 +08:00
lichenever 8854a6c9f7 fix two cast bug in auto parallel 2020-04-08 19:58:13 +08:00
fary86 73e0bdcd43 Dump graph with type info when static analysis failed 2020-04-08 19:58:13 +08:00
zhaozhenlong cca1d7764a add operator SpaceToBatch and BatchToSpace for ge 2020-04-08 19:58:13 +08:00
wenchunjiang 1c5f17ba3c This fixes an issue about mindspore process cannot exit when calling python api op_select_format failed in select kernel steps. 
Previously function op_select_format and check_supported raise an exception directly on the tbe_process python side, but we don't deal with the exception, and raise an exeception on c++ side to frontend ME, that will cause some conflict when recycle resource on ME and tbe_process python interpreter.
This changes adding try...catch in function op_select_format and check_supported on the python side, and return the Exception string to c++ side, so that we can raise an exception to frontend ME and ME will deal with resouce clearning and exit.
 2020-04-08 19:58:13 +08:00
panyifeng 33b92dddca fix parallel related valuenode merging error 2020-04-08 19:58:13 +08:00
c00425699 c4b03e854c use std::vector instead of std::list to promote performance for parallel module 2020-04-08 19:58:13 +08:00
lichenever 80b8d05c14 fix grpah mode loop sink bug in auto parallel 2020-04-08 19:58:13 +08:00
GinFung 2fd7f67108 Add matmul biasadd fusion pass 2020-04-08 19:58:13 +08:00
zhaozhenlong a6e626f02e add operator diag and diag_part 2020-04-08 19:58:13 +08:00
leonwanghui d2348c7f87 回退 'Pull Request !17 : [AutoParallel]Fix bug in the case of two cast' 2020-04-08 19:58:13 +08:00
lvliang c5da87254d auto enbale dynamic mem pool 2020-04-08 19:58:13 +08:00
lianliguang d3400cde01 fix bug of got a error transdata's dest format 2020-04-08 19:58:13 +08:00
shibeiji 2e308e4bed add log for kernel runtime in order to trace performance 2020-04-08 19:58:13 +08:00
lichenever c119f1c29e fix_cast_bug 2020-04-08 19:58:13 +08:00
buxue 24d1991cac Fix bug structure output when there is depend whose first input is constant in outputs 2020-04-08 19:58:13 +08:00
zhaozhenlong 04e4f6d927 adapt relu6grad and graphengine modified 2020-04-08 19:58:13 +08:00
mindspore-ci-bot b6d74f862e !178 fix compile error 
Merge pull request !178 from Simson/rebuild-graph-before-RunGraph
 2020-04-08 19:42:39 +08:00
jinyaohui ac62faa388 modify set_dataset_mode_config api param 2020-04-08 18:30:36 +08:00
simson a8bc8bfecb fix compile error 2020-04-08 17:48:42 +08:00
liuxiao 47d903ff57 Add pack and unpack 2020-04-08 16:58:53 +08:00
mindspore-ci-bot dd9a5a385a !175 Rebuild graph before RunGraph if needed 
Merge pull request !175 from Simson/rebuild-graph-before-RunGraph
 2020-04-08 16:58:38 +08:00
mindspore-ci-bot 2eb71103f9 !82 profiling feature enhancement 
Merge pull request !82 from caifubi/dev-profiling
 2020-04-08 16:22:35 +08:00
simson ee5b406b37 rebuild graph before rungraph if needed 2020-04-08 16:20:46 +08:00
mindspore-ci-bot 22516d3e08 !156 [Auto parallel] Separate memory_cost and computation_cost in cost model 
Merge pull request !156 from Xiaoda/implementing-memory-calculation-in-auto-parallel
 2020-04-08 16:11:49 +08:00
jonyguo 20d1b64443 fix: error info is not exactly when column list invalid 2020-04-08 16:11:13 +08:00
jojobugfree effdb483d6 profiling feature enhancement 2020-04-08 15:12:06 +08:00
anzhengqi fb4e84c0ee modify part of comments 2020-04-08 14:56:13 +08:00
Xiaoda Zhang a153fad874 This commit is to separate the computation cost and memory cost in auto_parallel. Some related memory correction is removed. 2020-04-08 11:52:19 +08:00
chenhaozhe d88dbbb138 pass auto mixed precision flag to ge init options 2020-04-08 11:43:46 +08:00
Jonathan Yan 9d0fde29f4 remove ENABLE_MINDRECORD flag 2020-04-07 14:55:45 -04:00
Alexey Shevlyakov 84d780c1a4 remove make_unique.h 2020-04-07 14:00:09 -04:00
mindspore-ci-bot 475e858474 !158 fix: resolve MindDataset hung when field not in index when using block_reader 
Merge pull request !158 from guozhijian/fix_block_reader_hung
 2020-04-07 21:49:33 +08:00
mindspore-ci-bot fe4c815d1c !151 updata mkl-dnn link and md5 
Merge pull request !151 from dengwentao/mkldnn_cmake
 2020-04-07 21:22:08 +08:00
jonyguo c688265671 fix: when use MindDataset block_reade=True hung 2020-04-07 17:53:23 +08:00
chenhaozhe b61ad0a5a7 use find instead of equal to distinguish training graph 2020-04-07 17:29:27 +08:00
mindspore-ci-bot 06e6c6b880 !128 refactor callback for ge backend 
Merge pull request !128 from wangqiuliang/refactor-call-back-for-ge
 2020-04-07 17:16:02 +08:00
dengwentao 2303719352 updata mkl-dnn link and md5 2020-04-07 16:27:44 +08:00
mindspore-ci-bot 30c242d70a !146 Fix some typo errors in mindspore ir module 
Merge pull request !146 from leonwanghui/typo-fix
 2020-04-07 14:40:06 +08:00
mindspore-ci-bot ece3e7921e !138 Change log level in pipeline_ge.cc 
Merge pull request !138 from chenhaozhe/change-log-level-in-pipeline-ge
 2020-04-07 10:55:12 +08:00
Wei Luning 113c0d8cd2 fix InsertGradientOf with class method 2020-04-07 02:11:35 +00:00
chenhaozhe e88bf9ca79 change log level in pipeline_ge 2020-04-07 10:04:11 +08:00
kingfo 3202fc0df9 refactor callback for ge backend 2020-04-07 09:32:54 +08:00
leonwanghui 322ffef3e4 Fix some typo errors in mindspore ir module 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-04-06 16:15:09 +08:00
mindspore-ci-bot 7a367af9c6 !135 fix grad missing due to indirect dependent free morphism 
Merge pull request !135 from penn/fix_free_morphism_error
 2020-04-06 10:17:13 +08:00
leonwanghui ad76e1c42a Fix some typo errors in pipeline module 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-04-04 13:00:24 +08:00
panyifeng 1fb776fe09 fix grad missing due to indirect dependent free morphism 2020-04-03 18:22:53 +08:00
wanghua 5b176f258b modify bert test file 2020-04-03 18:20:50 +08:00
mindspore-ci-bot 97276fa522 !86 Op Print: add comments and check function 
Merge pull request !86 from ghzl/improve-function-and-comment-in-op-print
 2020-04-03 17:46:42 +08:00
mindspore-ci-bot 9f982b513d !129 fix bert precision bug 
Merge pull request !129 from wanghua/master
 2020-04-03 17:35:44 +08:00
mindspore-ci-bot e579be472f !124 add suport for bool in print_kernels 
Merge pull request !124 from yankai10/add_support_for_bool
 2020-04-03 17:14:41 +08:00
mindspore-ci-bot d4e51c8f6e !116 Optimize operator information register 
Merge pull request !116 from zjun/master
 2020-04-03 16:43:47 +08:00
mindspore-ci-bot 2c3c1577b1 !115 Add parallel operator for ExpandDims 
Merge pull request !115 from yangzhenzhang/expand-dim
 2020-04-03 15:54:33 +08:00
zjun 651d9aae40 add new mode for operator info register 2020-04-03 15:54:12 +08:00
mindspore-ci-bot 02ffce800c !127 fix ME + GE compile error 
Merge pull request !127 from wangqiuliang/fix-ge-me-compile
 2020-04-03 15:46:11 +08:00
mindspore-ci-bot f639766f97 !11 Fix bug: Failed to get parameter data in model parallel scenes, when the size of tensor large than 2GB 
Merge pull request !11 from WeibiaoYu/fix-mem-problem
 2020-04-03 15:44:31 +08:00
guohongzilong 0eb72d76f0 import comment and function of op print 2020-04-03 15:41:09 +08:00
yangzhenzhang dd0d4e6b84 add parallel ops for expand dims 2020-04-03 15:17:07 +08:00
mindspore-ci-bot aba38a2401 !10 Add operator adapting in ME for SparseApplyFtrlD 
Merge pull request !10 from zhangzheng/SparseApplyFtrlD
 2020-04-03 15:13:28 +08:00
kingfo 01822c54d9 fix ME+GE compile error 2020-04-03 14:59:33 +08:00
wanghua da123c5b3e fix bert precison bug 2020-04-03 14:55:25 +08:00
mindspore-ci-bot d0e7ee38b4 !107 Add YOLOv3 infer scipt and change dataset to MindRecord 
Merge pull request !107 from zhaoting/add-YOLOv3-infer-scipt-and-change-dataset-to-MindRecord
 2020-04-03 14:38:28 +08:00
mindspore-ci-bot 367f513e8c !98 Pass str instead of bool to ge.exec.EnableDump 
Merge pull request !98 from ghzl/ge-exec-EnableDump-change-to-str
 2020-04-03 12:22:01 +08:00
mindspore-ci-bot b1bc4ea296 !117 new run package update 
Merge pull request !117 from jjfeing/master
 2020-04-03 12:02:50 +08:00
zhaoting 0c81759ae6 add YOLOv3 infer scipt and change dataset to MindRecord 2020-04-03 11:35:58 +08:00
y00445136 b3cd639c09 add support for bool 2020-04-03 11:05:38 +08:00
zhangz0911gm 96e1127edb Adding new operator SparseApplyFtrlD in ME 2020-04-02 22:28:21 -04:00
mindspore-ci-bot 3289d7bb69 !109 Fix some typo errors in session and device module 
Merge pull request !109 from leonwanghui/typo-fix
 2020-04-03 10:23:53 +08:00
jjfeing 8ece432ae9 tr5 run package update 2020-04-03 10:02:39 +08:00
yanghaoran b53c974513 modify reduceminD and reducemaxD IR 2020-04-02 22:10:03 +08:00
mindspore-ci-bot 9be1a01db6 !85 Optimize depend edge with make tuple input 
Merge pull request !85 from YuJianfeng/master
 2020-04-02 22:00:18 +08:00
mindspore-ci-bot a5a904fbdf !91 fix bug for allreduce fusion and add resnet unit test 
Merge pull request !91 from chentingting/allreduce_fusion_resnet
 2020-04-02 21:35:35 +08:00
mindspore-ci-bot ad88c55859 !105 register not equal distributed op 
Merge pull request !105 from lichen/register_not_equal_distributed_op
 2020-04-02 21:34:34 +08:00
mindspore-ci-bot cb323eccf0 !71 Add operator adapting in ME for FloorMod, Acosh and SparseApplyFtrlD 【CR】，并主动Acosh and FloorMod operator adapting 
Merge pull request !71 from zhangzheng/sparse_acosh_floormod
 2020-04-02 21:33:33 +08:00
mindspore-ci-bot a1d6f42be8 !28 fix custom op bug when custom op has muti-out and not use all of outputs 
Merge pull request !28 from geekun/master
 2020-04-02 21:02:25 +08:00
mindspore-ci-bot 55916351ee !52 remove ge depend 
Merge pull request !52 from flywind/remove-ge-depend
 2020-04-02 20:59:06 +08:00
zhangz0911gm 6f1747045c Add FloorMod, Acosh in ME 2020-04-02 08:50:39 -04:00
Wei Luning 73ba399364 remove ge depend in cpu 2020-04-02 19:38:42 +08:00
lichenever ff808021c7 register not equal distributed op 2020-04-02 19:28:30 +08:00
c00425699 ab917a734d fix bug for allreduce fusion and add resnet unit test 2020-04-02 18:55:04 +08:00
mindspore-ci-bot d8b460c780 !96 fix refkey bug for auto parallel 
Merge pull request !96 from lichen/fix_ref_key_bug_for_auto_parallel
 2020-04-02 17:26:58 +08:00
leonwanghui fdb47860cc Fix some typo errors in session and device module 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-04-02 17:02:52 +08:00
lichenever 5240b1f603 fix refkey bug for auto parallel 2020-04-02 16:14:29 +08:00
WeibiaoYu b78427e619 the size of tensor may be bigger than 2GB, should use memcpy instead of memcpy_s 2020-04-02 03:52:29 -04:00
Su Teng 60b68a1470 sort include file in parallel dir 2020-04-02 15:14:47 +08:00
geekun c42e53ae5b fix custom op bug and add custom op check 2020-04-02 14:28:37 +08:00
mindspore-ci-bot 5acb93a345 !79 Dump Failed when trans_flag set to true 
Merge pull request !79 from laiyongqiang/develop
 2020-04-02 14:13:42 +08:00
guohongzilong 15d061f852 pass str to ge.exec.EnableDump 2020-04-02 12:05:28 +08:00
YuJianfeng 7f53bb062d Optimize depend edge with make tuple input 2020-04-02 11:33:34 +08:00
mindspore-ci-bot cf54ecfe6e !62 support grad on python function with variable arguments 
Merge pull request !62 from amongo/SupportGradOnVarArgs
 2020-04-02 11:21:06 +08:00
mindspore-ci-bot 85df19b23c !66 Add RMSProp optimizer 
Merge pull request !66 from zhaoting/add-RMSProp
 2020-04-02 09:57:13 +08:00
mindspore-ci-bot a44b5293de !68 fix bug: Failed to get parameter data in model parallel scenes, when the size of tensor large than 2GB 
Merge pull request !68 from WeibiaoYu/master
 2020-04-02 09:33:12 +08:00
zhaoting ed3c2d7229 add RMSProp optimizer 2020-04-02 09:10:58 +08:00
mindspore-ci-bot f95a5b9027 !81 [Auto parallel] Making non-identity triangle elimination exact in DP algorithm 
Merge pull request !81 from Xiaoda/change_star_elimination-make_the_non-identity_triangle_eliminatin_exact
 2020-04-02 09:04:03 +08:00
mindspore-ci-bot 7341d7ec1e !83 add get ptr for memreuseutil 
Merge pull request !83 from kisnwang/add-getptr-for-memreuseutil
 2020-04-01 22:08:52 +08:00
kswang 04be6a37f0 add getptr for memreuse 2020-04-01 21:37:02 +08:00
huangdongrun d3f733fa25 support grad on graph with variable arguments 
use unpack graph primitive instead

add testcases for all grad interface

remove debug log

format code

remove dumpfuncgraph

resolve clang-format

resolve reviews

resolve cpplint

fix review
 2020-04-01 20:24:04 +08:00
Xiaoda Zhang 3d35792877 change_star_elimination: make the non-identity triangle_eliminatin exact 2020-04-01 19:24:18 +08:00
mindspore-ci-bot 69d5403319 !14 add adapter for ge atan2 
Merge pull request !14 from zhaozhenlong/op/atan2
 2020-04-01 18:24:55 +08:00
mindspore-ci-bot 0f2ed0b134 !61 Modify log level in DfGraphManager 
Merge pull request !61 from chenhaozhe/modify-log-info-of-dfgraphmanager
 2020-04-01 18:06:41 +08:00
zhaozhenlong bed5997a14 add atan2 operator 2020-04-01 16:27:35 +08:00
laiyongqiang bb421b70db fix bug when dump op with trans_flag true 2020-04-01 16:21:00 +08:00
mindspore-ci-bot b829be11a6 !37 Quick fix for pre-opensource TFReaderOp issue 
Merge pull request !37 from Peilin/peilin-pre-opensource-tfreader-fix
 2020-04-01 15:32:08 +08:00
mindspore-ci-bot 22a9c00bcd !57 Add parallel operators for Neg and BatchMatMul 
Merge pull request !57 from yangzhenzhang/master
 2020-04-01 15:23:46 +08:00
mindspore-ci-bot 569e5b056e !53 [Auto parallel] Remove the redundant work in star elimination 
Merge pull request !53 from Xiaoda/modify_star_and_triangle_elimination
 2020-04-01 11:43:31 +08:00
mindspore-ci-bot 8f76c28017 !56 remove cce dependency 
Merge pull request !56 from jojobugfree/dev-rm-cce
 2020-04-01 10:12:54 +08:00
mindspore-ci-bot 87040483ee !58 fix two cast bug in auto parallel 
Merge pull request !58 from lichen/fix_two_cast_bug_in_auto_parallel
 2020-04-01 09:54:10 +08:00
WeibiaoYu 60f7a95b1c the size of tensor may be bigger than 2GB, should use memcpy instead of memcpy_s 2020-03-31 21:47:28 -04:00
yangzhenzhang 110640e2ad add parallel ops for neg and batchmatmul 2020-04-01 09:36:38 +08:00
Xiaoda Zhang c080ec7874 change star elimination: remove some redundant and checking works 2020-04-01 09:16:36 +08:00
Peilin Wang 0ae77bb0db TFReaderOp fix, threads will exit after reading necessary amount of rows 
changes from yanpanhui 524009: added set_dataset_size and changed get_dataest_size according to ME requirements

CI fixes
 2020-03-31 13:29:54 -04:00
mindspore-ci-bot 52166a85cf !43 dump graph with type info when analysis failed 
Merge pull request !43 from fary86/dump-typed-graph-when-analyze-fail
 2020-03-31 22:23:02 +08:00
mindspore-ci-bot 045a1c356c !51 add op SpaceToBatch and BatchToSpace for ge 
Merge pull request !51 from zhaozhenlong/op/space-to-batch
 2020-03-31 21:59:03 +08:00
chenhaozhe 0727c2e76a modify log level in DfGraphManager 2020-03-31 21:09:35 +08:00
fary86 816b60491d Dump graph with type info when static analysis failed 2020-03-31 21:08:48 +08:00
mxm 2f031e1540 fixed: PrimitiveToInferImplMap map is global, and key of the map PrimitivePtr also a global variable. If key is initialized later than the map initialized during compilation, will cause the primitive map initialize failed. Variable initialization order is not guaranteed during compilation. 2020-03-31 20:49:33 +08:00
mindspore-ci-bot a8450dd761 !47 fix issue about the process don't exit when custom ops run failed 
Merge pull request !47 from wchunjiang/master
 2020-03-31 20:28:05 +08:00
zhaozhenlong cf40305bf0 add operator SpaceToBatch and BatchToSpace for ge 2020-03-31 20:21:47 +08:00
mindspore-ci-bot da4c711dfb !50 fix parallel related valuenode merging error 
Merge pull request !50 from penn/fix_parallel_related_valuenode_merging_error
 2020-03-31 19:44:37 +08:00
panyifeng feb1c36811 fix parallel related valuenode merging error 2020-03-31 19:40:16 +08:00
lichenever 2da38ad401 fix two cast bug in auto parallel 2020-03-31 18:54:21 +08:00
mindspore-ci-bot da447b8d4d !45 use std::vector instead of std::list to promote performance for parallel module 
Merge pull request !45 from chentingting/permote_auto_prallel_performance
 2020-03-31 18:25:59 +08:00
jojobugfree e6c15b82c8 remove cce dependency 2020-03-31 17:02:25 +08:00
lichenever f946aea10d fix grpah mode loop sink bug in auto parallel 2020-03-31 16:16:53 +08:00
wenchunjiang 5a00d8cb58 This fixes an issue about mindspore process cannot exit when calling python api op_select_format failed in select kernel steps. 
Previously function op_select_format and check_supported raise an exception directly on the tbe_process python side, but we don't deal with the exception, and raise an exeception on c++ side to frontend ME, that will cause some conflict when recycle resource on ME and tbe_process python interpreter.
This changes adding try...catch in function op_select_format and check_supported on the python side, and return the Exception string to c++ side, so that we can raise an exception to frontend ME and ME will deal with resouce clearning and exit.
 2020-03-31 15:46:40 +08:00
c00425699 3bb48ffee1 use std::vector instead of std::list to promote performance for parallel module 2020-03-31 15:16:00 +08:00
mindspore-ci-bot 976226f9ac !10 Add matmul biasadd fusion pass 
Merge pull request !10 from YuJianfeng/master
 2020-03-31 14:52:47 +08:00
mindspore-ci-bot 6f03881b04 !39 add op Diag and DiagPart 
Merge pull request !39 from zhaozhenlong/op/diag
 2020-03-31 14:45:44 +08:00
mindspore-ci-bot 7242c0fad5 !41 Revert 'Pull Request !17 : [AutoParallel]Fix bug in the case of two cast' 
Merge pull request !41 from leonwanghui/revert-merge-17-master
 2020-03-31 14:18:17 +08:00
mindspore-ci-bot e4b404e8ae !32 auto-enable-dynamic-mem-pool 
Merge pull request !32 from JoyLvliang/master
 2020-03-31 14:12:24 +08:00
mindspore-ci-bot e535a5f50c !40 fix bug of get a error dst format of transdata 
Merge pull request !40 from lianliguang/master
 2020-03-31 11:59:10 +08:00
mindspore-ci-bot 84a61bd015 !27 add log for kernel runtime in order to trace performance 
Merge pull request !27 from shibeiji/master
 2020-03-31 11:13:27 +08:00
lvliang b3a306489d auto enbale dynamic mem pool 2020-03-31 10:30:01 +08:00
leonwanghui 976af212e9 回退 'Pull Request !17 : [AutoParallel]Fix bug in the case of two cast' 2020-03-31 10:27:40 +08:00
mindspore-ci-bot 140a15924c !17 [AutoParallel]Fix bug in the case of two cast 
Merge pull request !17 from lichen/fix_cast_bug
 2020-03-31 10:25:04 +08:00
mindspore-ci-bot 02a25407c4 !30 use string::find instead of equal to distinguish training graph 
Merge pull request !30 from chenhaozhe/enable-variable-acc-for-training-graph
 2020-03-31 10:18:06 +08:00
lianliguang 9d5890d9b9 fix bug of got a error transdata's dest format 2020-03-31 10:04:34 +08:00
zhaozhenlong b12e6ff780 add operator diag and diag_part 2020-03-31 09:44:56 +08:00
mindspore-ci-bot c1c8fef9ca !24 Change strategy for structure output 
Merge pull request !24 from 步学/structure-output
 2020-03-31 09:38:44 +08:00
lichenever b4d34973bc fix_cast_bug 2020-03-30 21:13:45 +08:00
chenhaozhe cab5503280 use find instead of equal to distinguish training graph 2020-03-30 20:10:56 +08:00
buxue 0da0bdcf40 Fix bug structure output when there is depend whose first input is constant in outputs 2020-03-30 19:49:46 +08:00
zhaozhenlong 9862dea3cf adapt relu6grad and graphengine modified 2020-03-30 19:01:03 +08:00
shibeiji 468e257a14 add log for kernel runtime in order to trace performance 2020-03-30 17:55:28 +08:00
GinFung 468dbc3557 Add matmul biasadd fusion pass 2020-03-30 10:12:50 +08:00
zhunaipan 930a1fb0a8 initial version 
Signed-off-by: leonwanghui <leon.wanghui@huawei.com>
 2020-03-27 22:54:54 +08:00