mindspore-Ecosystem
/
mindspore
21
0
Fork 27
Code Issues Pull Requests 6 Projects Releases Wiki Activity
62,951 Commits 23 Branches 30 Tags 4 GiB
Commit Graph

145 Commits

Author SHA1 Message Date
i-robot 960f5efffe
!49942 move head files to include 
Merge pull request !49942 from zhoufeng/move-head-files-to-include
 2023-03-08 08:45:15 +00:00
zhoufeng b7bc99debf move head files to include 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2023-03-08 10:44:43 +08:00
reku1997 5b91ce9b64 fix convert dynamic broadcast to pass 2023-03-06 15:10:46 +08:00
ttudu a545cb1d84 change kernel_build_info default to unknown 2023-02-08 11:33:31 +08:00
ttudu 4f4f1d1cf1 ascend ops support backoff 2023-02-07 10:42:40 +08:00
ttudu 37586dbf9b ascend support untuple 2023-01-31 20:38:28 +08:00
i-robot 923d05c0e2
!46459 delete useless max/min shape 
Merge pull request !46459 from chengbin/delete_min_max
 2022-12-12 01:15:51 +00:00
ckey_Dou 78b83a1a54 delete useless min/max shape 2022-12-11 16:06:34 +08:00
jjfeing 82f04825f3 remove enum fusion type 2022-12-08 11:19:48 +08:00
reku1997 d84c34c795 fix dropout domask format select 2022-11-30 10:22:27 +08:00
jjfeing 90a3865918 dynamic static op adapter 2022-11-22 16:13:04 +08:00
luoyuan c0c76f7f17 [MS][OPS] unsorted_segment_sum op dynamic compile static and add irfission 2022-09-29 20:24:40 +08:00
linqingke 8ec70d2f14 Fix bert and transformer performance. 2022-09-15 15:44:22 +08:00
LaiYongqiang 03d1165618 add attr rename and op type rename for ascend vm op adapter 2022-09-05 10:18:25 +08:00
LaiYongqiang e809917a50 move ascend const input to attr pass to unify mindir 2022-08-27 17:21:00 +08:00
i-robot b06f5623de
!40497 new convert op info register 
Merge pull request !40497 from laiyongqiang/new_reg
 2022-08-22 11:37:17 +00:00
LaiYongqiang 69355958d7 new convert op info register 2022-08-18 21:16:03 +08:00
LaiYongqiang 1af40c4d0a refactory check support functions 2022-08-18 16:46:56 +08:00
lby 1aa76afba5 decoupling plugin and runtime 2022-07-20 10:48:18 +08:00
LaiYongqiang 1e95729ff1 create new cnode when node is changed 2022-07-15 15:22:08 +08:00
muxiyin 27e96dd46d support spmm 2022-06-13 15:08:09 +08:00
yiyanzhi_akane 1e18fe3574 AdamApplyOneWithDecay support dynamic shape 2022-06-08 09:53:01 +08:00
huoxinyou e7c769eebd AdamApplyOne support dynamic shape & 
ScatterNd ops's indices support int64 docs
 2022-06-02 20:21:33 +08:00
Zhang Qinghua a89b662b62 Change the weight parameter to FV parameter in FuncGraph, and add ST cases. 2022-05-31 14:27:15 +08:00
ttudu ab8b8775b5 delete updateop actor 2022-04-20 09:44:11 +08:00
yuchaojie bb52ffb2a9 add TensorScatter op fission 2022-04-16 14:57:04 +08:00
ttudu adb18b82ca dynamic shape CPU & GPU 2022-03-22 13:03:31 +08:00
zhoufeng b7bb53ff73 build libmindspore_backend.so 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2022-03-17 19:14:45 +08:00
zhoufeng f49b195c39 extract common as an independent shared library 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2022-02-26 10:55:51 +08:00
chenlei_autodiff 5df9f01310 [GraphKernel] decouple convert_const_input_to_attr. 2022-02-15 10:27:34 +08:00
marui d47e79b04c Refactor ccsrc directories and CMakeLists files 2022-02-13 14:37:51 +08:00
huanghui f3ebb1a0e1 adapt some cpp ut for unused nodes eliminate 2022-01-21 11:30:39 +08:00
TronZhang 58f386fe2a Add Custom actor nodes and custom actor support for dynamic shape case 2022-01-07 17:31:37 +08:00
zhoufeng 9d432d029d modify dev env name 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2021-12-16 14:12:24 +08:00
He Wei 41dcac9c49 Replace std::unordered_map/set with robin-hood-hashing 
Robin-hood-hashing (https://github.com/martinus/robin-hood-hashing)
is considered faster then std::unordered_map/set,
so we use it to improve mindspore performance.

1. robin_hood head file in `third_party/robin_hood/include`;
2. In `utils/hash_map.h` and `utils/hash_set.h`, we define:
 - mindspore::HashMap as an alias of robin_hood::unordered_map;
 - mindspore::HashSet as an alias of robin_hood::unordered_set;
3. Replace:
 - `#include <unordered_map>` --> `#include "utils/hash_map.h"`;
 - `#include <unordered_set>` --> `#include "utils/hash_set.h"`;
 - `std::unordered_map` --> `mindspore::HashMap`;
 - `std::unordered_set` --> `mindspore::HashSet`;
 - `map.insert(std::pair(key, value))` --> `map.emplace(key, value)`;
 - `[] (const std::pair<K, V> &p) {..} ` --> `[] (const auto &p) {..} `;
4. Fix issues found by switch to robin_hood:
 - AnfNodeConfig hash and equal;
 - Fix a bug in `Slice::operator==()`;
 - Fix a bug in `CNode::HasPrimalAttr()`;
 - Fix map.erase() usage bugs: `map.erase(iter++)` --> `iter = map.erase(iter)`;
 - Fix some iterator invalidated problem;
5. Some std::unordered_map/set can not replace by robin_hood:
 - As parameter of functions that exposed to python by pybind11;
 - Use bad hash that cause robin_hood::map over_flow, such as AbstractBasePtrListHasher;
6. Update cpp unit tests;
7. Add build option '-F' to enable robin_hood, default on.
 2021-11-24 10:47:40 +08:00
dayschan cbb84ff580 Move IsRealKernel and IsRealCNodeKernel from AnfAlgo to AnfUtils 
the function IsOneOfPrimitive and IsOneOfPrimitiveCNode is useful,
we can move them into anf.cc
 2021-11-09 14:34:21 +08:00
zhoufeng fb17c448c2 add pass switch 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2021-11-01 16:33:02 +08:00
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
zhoufeng 03a56f2bb0 alltoall exception handle 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2021-08-11 17:30:49 +08:00
zhoufeng 2143241092 change neighbor exchange to all to all 
Signed-off-by: zhoufeng <zhoufeng54@huawei.com>
 2021-08-01 18:21:28 +08:00
laiyongqiang 1533435015 replace memcpy_async with tensor move 2021-04-15 16:12:01 +08:00
dingpeifei 3c9d8cb073 The input and output of batchnorm reverse operator increase pass in ascend platform under the mode of pynitve 2021-03-27 16:49:38 +08:00
dingpeifei 87e41aaeee IR operators of GPU and CPU are unified as batchnorm 2021-03-18 19:02:28 +08:00
liubuyu 518818fbef reshape type for 3d nodes 2021-03-10 08:56:10 +08:00
yuchaojie 6d195f340c add SyncBatchNorm 2021-02-27 14:32:54 +08:00
mindspore-ci-bot aebe263dce !11895 unify mindir for different backend: the output num of optimizer ops, the backward of concat 
From: @wangnan39
Reviewed-by: 
Signed-off-by:
 2021-02-05 16:27:19 +08:00
jinyaohui 8022f9a6ed modify pack to stack 2021-02-04 18:54:31 +08:00
wangnan39@huawei.com cd9173fdfd unify the output num of optimizer ops 2021-02-04 17:46:00 +08:00
shenghong96 49144fde37 fix UT of test_topk_no_split 2021-02-02 16:52:20 +08:00