p60975123
/
mindspore
forked from mindspore-Ecosystem/mindspore
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
24,091 Commits 23 Branches 0 Tags 3.9 GiB
Commit Graph

24091 Commits

Author SHA1 Message Date
zhanghuiyao 07ee4054c5 fix some network built-in function problem. 2021-05-28 16:58:10 +08:00
mindspore-ci-bot 56bb6468df !16800 covert python to C++ in BiasAddGrad operator 
From: @shen_jingxing
Reviewed-by: @ginfung,@zh_qh
Signed-off-by: @zh_qh
 2021-05-28 16:57:10 +08:00
mindspore-ci-bot bfeb511100 !16967 clean code check 
From: @Margaret_wangrui
Reviewed-by: @ginfung
Signed-off-by:
 2021-05-28 16:55:00 +08:00
mindspore-ci-bot 36d3c59c23 !17108 clean pclint check 
From: @Margaret_wangrui
Reviewed-by: @ginfung,@zh_qh
Signed-off-by: @zh_qh
 2021-05-28 16:54:50 +08:00
zhaoting e938898336 add fp16 support for CPU Adam 2021-05-28 16:52:08 +08:00
累到崴脚 59935cd7d9 update 
updaterun_eval.sh

'update'

'update'

'update'

'updatereadme'

'fix'

update eval.py

'update'

update README

fixREADME

fixREADME

fix_help

to_research

fix

fixnetwork

fix

2021

fix

fix

fix
 2021-05-28 16:45:09 +08:00
mindspore-ci-bot 0d8ec8db88 !17143 clean code of probability 
From: @bingyaweng
Reviewed-by: @wang_zi_dong,@sunnybeike
Signed-off-by: @sunnybeike
 2021-05-28 16:27:47 +08:00
mindspore-ci-bot f96e149e1e !16977 ascend 310&910 joint compile, and ascend 310 package lose weight 
From: @nicholas_yhr
Reviewed-by: @zhoufeng54,@zhoufeng54,@xsmq
Signed-off-by: @zhoufeng54,@xsmq
 2021-05-28 16:18:45 +08:00
zengzitao 07752d7aaf fix warpctc bug when open graph_kernel flag 2021-05-28 16:14:49 +08:00
huangbo77 eaadd8746a fixes improper usage of built-in function _Loss 2021-05-28 16:13:25 +08:00
mindspore-ci-bot ea93cc380a !17000 clear code check warnings 
From: @shibeiji
Reviewed-by: @wuxuejian,@wuxuejian,@liangchenghui,@wuxuejian
Signed-off-by: @wuxuejian,@liangchenghui
 2021-05-28 16:04:36 +08:00
mindspore-ci-bot ae5adc2986 !17155 add st model of micro 
From: @yangjie159
Reviewed-by: @wangchengyuan,@hangangqiang
Signed-off-by: @wangchengyuan,@hangangqiang
 2021-05-28 15:31:51 +08:00
mindspore-ci-bot ecbdb6a853 !17063 maskrcnn_mobilenetv1 export 
From: @zhangxiaoxiao16
Reviewed-by: @c_34,@oacjiewen
Signed-off-by: @c_34
 2021-05-28 15:29:59 +08:00
mindspore-ci-bot f5fb195f04 !17189 check input model buffer size 
From: @hangangqiang
Reviewed-by: @zhanghaibo5,@zhang_xue_tong
Signed-off-by: @zhanghaibo5,@zhang_xue_tong
 2021-05-28 15:17:54 +08:00
VectorSL 03210aee81 clean code2 2021-05-28 15:15:45 +08:00
YangLuo ac319b3a79 Fix mindrecord UTs: files existed causes write exception 2021-05-28 15:13:48 +08:00
caifubi 8056d9709c fix gpu pynative allreduce fail 2021-05-28 15:12:00 +08:00
Su Teng 54eeb3e789 Add sparse attenton related ops 2021-05-28 15:10:40 +08:00
mindspore-ci-bot 40ca285ab3 !17052 added scale out 
From: @anancds
Reviewed-by: @cristoval,@limingqi107
Signed-off-by:
 2021-05-28 15:05:39 +08:00
wangshuide2020 0f9745bd20 add default value of optional attribute for nllloss. 2021-05-28 14:56:50 +08:00
dayschan 2ac8c65327 Add GraphKernelPassManager to manage the passes of GraphKernel 
Refactor the original "PassManager" class, and derive the "GraphKernelPassManager"
GraphKernel's ir files are dumped into a new sub-directory "graph_kernel" in the original "verbose_ir_files"

All GraphKernel's passes are divided into 3 levels, and controlled by the flag "opt_level" by default.
when the opt_level is greaterequal to the pass's level, this pass will run.
The default "opt_level" is 2 when GraphKernel is enabled.
Levels:
  1. Basic features, like cluster, splitter, and some preprocess, postprocess.
  2. All stable features, mainly includes the optimization passes.
  3. Experimental features, like stitch-fusion, parallel-fusion.

The two flags "enable_pass" and "disable_pass" are available in this commit.
User can manually enable some passes when it's disabled by "opt_level", or disable the enabled passes,
by specifying that pass in this format: "stage_id.pass_id" or "stage_name.pass_name", multiple passes are separated by comma(",")
the stage/pass index and stage/pass name can be found from the ir filename.
e.g.   "--enable_pass=cluster.graph_kernel_expander,1.1,1.2"

Others:
  1. the pass "tensor_promotion" is not useful, remove it.
  2. put the pass "InsertPadOps" before "ArithmeticSimplify".
 2021-05-28 14:50:37 +08:00
ling 58d30d0c65 mindrt modify include .h 
device_type
 2021-05-28 14:42:46 +08:00
mindspore-ci-bot 9dcc495c24 !17109 [MD] fix code check 
From: @liyong126
Reviewed-by: @liucunwei,@jonyguo
Signed-off-by: @liucunwei
 2021-05-28 14:32:09 +08:00
simson 8d701f17d7 add check of undetermined type in addn infer 2021-05-28 14:27:07 +08:00
mindspore-ci-bot 30be164856 !16767 ci format checking issues fix 
From: @zhangzhaoju
Reviewed-by: @hwhewei,@ginfung,@zh_qh
Signed-off-by: @zh_qh
 2021-05-28 14:16:48 +08:00
mindspore-ci-bot 932ec6e643 !17034 revert convback and add convtranspore 
From: @changzherui
Reviewed-by: @zhoufeng54,@liangchenghui
Signed-off-by: @liangchenghui
 2021-05-28 14:16:41 +08:00
zuochuanyong 8302820597 switch implementation of BiasAdd op to nnacl 2021-05-28 14:07:10 +08:00
mindspore-ci-bot 14cf33a6df !17197 回退 'Pull Request !17026 : add relative-position-attention in mindspore lite & nnacl' 
From: @xsmq
Reviewed-by: @zhoufeng54,@guoqi1024
Signed-off-by: @guoqi1024
 2021-05-28 13:42:04 +08:00
gongxiaoqing f26891df3a 回退 'Pull Request !17026 : add relative-position-attention in mindspore lite & nnacl' 2021-05-28 12:53:40 +08:00
liuyihong 143009a51b enable graph kernel 2021-05-28 12:40:03 +08:00
yanghaoran 8b4ff8708f ascend 310/910 joint compile, and ascend 310 pacakge lose weight 2021-05-28 12:11:20 +08:00
hangangqiang 8e1b5f8220 check input model buffer size 2021-05-28 11:39:33 +08:00
looop5 68f55e1e93 expand conv2d when input format is DefaultFormat but attr format is NHWC 2021-05-28 11:38:32 +08:00
wanyiming 6dfb903988 clean cpu code 2021-05-28 11:37:16 +08:00
mindspore-ci-bot 9fdc0035a0 !17046 add modelzoo network retinaface_resnet50 testcase 
From: @anzhengqi
Reviewed-by: @c_34,@wuxuejian
Signed-off-by: @c_34
 2021-05-28 11:09:58 +08:00
yuzhenhua 1a69ef0a3c ascend310 inference for xception 2021-05-28 11:09:37 +08:00
mindspore-ci-bot 63ba14b5c6 !17026 add relative-position-attention in mindspore lite & nnacl 
From: @hangangqiang
Reviewed-by: @zhanghaibo5,@zhang_xue_tong
Signed-off-by: @zhang_xue_tong
 2021-05-28 10:54:24 +08:00
baihuawei 8665faac0a clear warning 2021-05-28 10:43:29 +08:00
yangjie159 5cf7037ecd add st models 2021-05-28 10:05:59 +08:00
mindspore-ci-bot f5a23ddf26 !17168 drop test_kron 
From: @jachua
Reviewed-by: @wmzheng2020,@xsmq
Signed-off-by: @xsmq
 2021-05-28 10:01:01 +08:00
huangmengxi 0a588f1f44 drop test kron 2021-05-28 09:44:43 +08:00
mindspore-ci-bot 615a22b071 !16791 Comprehensive API changes part 2 
From: @dinglinhe123
Reviewed-by: @liangchenghui,@liangchenghui
Signed-off-by: @liangchenghui,@liangchenghui
 2021-05-28 09:37:59 +08:00
Yang Jiao f9aefb24d5 enable tinybert graph kernel 2021-05-28 09:35:36 +08:00
shen_jingxing 87175b9d22 BiasAddGrad 2021-05-28 09:22:50 +08:00
Ziyan 4b17493e52 handle load in step parallel 2021-05-28 09:16:40 +08:00
liyong 322c342979 fix codecheck 2021-05-28 09:14:29 +08:00
zhangxinfeng3 d86465602a clean code of probability 2021-05-28 08:58:40 +08:00
hangangqiang 1329355b33 add relative-position-attention in mindspore lite & nnacl 2021-05-28 08:52:32 +08:00
changzherui d9e2da299d Revert "!16599 c++ infer for conv2dbackpropfilter and conv2dbackpropinput" 
This reverts commit 3be79efd80, reversing
changes made to cf4479756a.
 2021-05-27 22:24:13 +08:00
mindspore-ci-bot f22e0522fe !16947 fix precision error of ml_location_lane_counter.onnx 
From: @wangyanling10
Reviewed-by: @zhanghaibo5,@jpc_chenjianping
Signed-off-by: @jpc_chenjianping
 2021-05-27 21:16:20 +08:00