a18949446876
/
mindspore2022
forked from huawei/mindspore2022
0
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
34,388 Commits 18 Branches 22 Tags 3.5 GiB
Commit Graph

55 Commits

Author SHA1 Message Date
lilei f8827a09fb modify parallel API note for master 2021-11-24 17:44:35 +08:00
huangxinjing f354ab22a3 add pipeline shard interface 
Add support for no pipeline accugradient

Add delay tag for fusion op

Optimizer the visite order

add mirror for mini step control

Move the group to attributes

Add gradient_shard control for the mini step

Fix code stype

Fix ut description

Add interface
 2021-11-09 09:11:13 +08:00
i-robot b8ec196d6e !25708 Add mode check 
Merge pull request !25708 from huangxinjing/add_mode_check
 2021-11-04 07:09:06 +00:00
i-robot 27be612516 !25719 Add Device number check 
Merge pull request !25719 from huangxinjing/add_device_number_check
 2021-11-04 01:34:22 +00:00
huangxinjing bd5cfafe2c Add mode check 
Add parallel mode check
 2021-11-04 09:05:34 +08:00
huangxinjing 741e3604eb Add Device Number Check 2021-11-01 15:36:50 +08:00
cononlly cc83b8064f Fix code for GPU Pipeline 2021-10-26 10:41:51 +08:00
lichenever 56b31fce42 add_no_elimilate_for_comm_op 2021-09-03 16:54:56 +08:00
i-robot 04c2e928c4 !21622 modify code check for Master 
Merge pull request !21622 from lilei/modify_code_check_master
 2021-08-26 12:18:11 +00:00
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
 2021-08-23 01:46:38 +08:00
djc 4e6f7dc97d [feat] [assistant] [I3T96X] add new Dataset operator LibriSpeechDataset 2021-08-22 13:39:37 +08:00
lilei 5714d68d1d modify code check for master 2021-08-16 15:58:15 +08:00
yao_yf a83bf73298 union auto_parallel_context interface dataset_strategy 2021-08-16 10:54:46 +08:00
yao_yf dc7dc7d3fa dataset strategy set 2021-07-27 18:44:10 +08:00
Xiaoda Zhang bb5d4212f7 enable All2All in infering redistribution ops 2021-07-24 09:09:01 +08:00
Xiaoda Zhang 04381273b3 Add the sharding propagation function: 
1) users configure sharding strategies for operators;
2) framework will propagate the strategies from configured-ops to
non-configured ops using BFS;
3) the propagation goal is to minimize redistribution communication
cost;
 2021-07-14 10:54:18 +08:00
lichenever cb438ce350 rectification_log 2021-06-17 09:23:17 +08:00
Ziyan 95ac0f6d58 fix optimizer weight shard config 2021-06-08 10:44:17 +08:00
huangxinjing e79db658e8 Fix codex for python file 2021-05-21 08:57:35 +08:00
Ziyan 2a752f24bf enable not fully use opt shard 2021-05-07 15:32:48 +08:00
Ziyan d19d42ee44 modify grad accu and comm fusion api 2021-03-19 15:11:40 +08:00
liujunzhu 6541b96c40 Add communication parallel mode. 2021-03-05 21:36:03 +08:00
yangzhenzhang 7303c3d3b8 add group ckpt 2021-01-21 14:41:05 +08:00
yangzhenzhang 9da3f9bec9 mini step grad accumulation 2021-01-06 09:25:06 +08:00
lizhenyu 7eb49cfce7 [bugfix] server core dump after traning 2020-12-28 20:25:45 +08:00
jinyaohui e6f9806cfb add broadcast 2020-11-18 09:32:27 +08:00
lichenever cfffff2875 add check for allreduce fusion 2020-10-12 16:55:30 +08:00
huangxinjing 2fa6a3b3c2 Fix doc error 2020-09-28 19:17:26 +08:00
mindspore-ci-bot 9bd34a1b29 !6673 Add stage information for ops and strategy 
Merge pull request !6673 from huangxinjing/stage_strategy
 2020-09-27 11:31:07 +08:00
huangxinjing 4ef439e27b Add stage information for ops and strategy 2020-09-25 10:42:43 +08:00
lichenever 395d3f0848 add_limit_for_allreduce_fusion 2020-09-24 15:06:58 +08:00
huangxinjing 8ba1503135 Add default value for auto search parallel mode 2020-09-18 19:44:30 +08:00
yao_yf b70204c080 auto parallel context add notes and func mv 2020-09-17 21:09:11 +08:00
Ziyan 8ea177e614 fix_api_problems 2020-09-17 11:32:13 +08:00
lichenever f2d3fd34ce rectification_allreduce_fusion_api 2020-09-07 16:38:58 +08:00
yao_yf d4cfe55c04 rename mirror_mean to gradients_mean 2020-09-04 14:47:36 +08:00
yao_yf 8f7aa5bd5a auto parallel context modify 2020-08-31 21:19:49 +08:00
Yi Huaijie 394be43492 raise RuntimeError when set different mode after Initializer created 2020-08-24 14:08:09 +08:00
Yi Huaijie 89a4ebf8a1 parallel mode must be set before create an initializer 2020-08-20 19:20:46 +08:00
yuchaojie 64a1560f1a add allreduce group for resnet gpu version 2020-08-11 17:26:39 +08:00
yuchaojie ed9cf2036c add nccl default allreduce fusion group 2020-07-28 09:47:36 +08:00
Ziyan 39f08eb7dd enable optimizer parallel 2020-07-16 11:47:41 +08:00
Ziyan 0925e35252 enable optimizer parallel with broadcast 2020-06-27 14:36:17 +08:00
hongxing d798325127 supplement description about default value 2020-06-20 08:28:51 +01:00
hongxing 3ad3a71fc7 change interface 2020-06-19 00:08:55 +01:00
mindspore-ci-bot 4df861cb62 !1672 support load full dataset on each device 
Merge pull request !1672 from yihuaijie/dev
 2020-06-12 09:24:04 +08:00
kswang 699166e552 default fusion group for ge 2020-06-11 20:44:24 +08:00
Yi Huaijie e5c351690b support load full dataset on each device 2020-06-11 19:48:06 +08:00
kswang 503dd297c5 set all reduce fusion default group 2020-05-19 21:29:07 +08:00
kswang 362bbacf19 add group for allreduce fusion 2020-05-18 10:59:08 +08:00