Commit Graph

34755 Commits

Author SHA1 Message Date
i-robot 95813a0b0a !26866 deal with kernel_graph_#_ prefix in tensor name in overflow wp checking
Merge pull request !26866 from john_tzanakakis/jt_bug_fixes
2021-11-27 09:16:02 +00:00
i-robot 8af53e9ae5 !26884 update api cn format
Merge pull request !26884 from yingchen/code_docs_zpi1127
2021-11-27 09:01:02 +00:00
zhi-qing-wei e3623a25f2 [feat] [assistant] [I48OCF] add dynamic shape for ReciprocalGrad operator 2021-11-27 16:53:48 +08:00
yingchen 319c67df14 update api format 2021-11-27 16:44:36 +08:00
wang_shaocong 2b356dd4a0 [MSLITE] Fix ci smoketest 2021-11-27 16:21:28 +08:00
ling 0d7c92b750 [MSLITE] create lite session failed bug 2021-11-27 16:14:58 +08:00
i-robot 591e606de9 !26873 move graph_kernel_cluster to directory graph_kernel/core
Merge pull request !26873 from DeshiChen/1127_cluster
2021-11-27 08:05:31 +00:00
yujianfeng 8bad0f581c Remove the hash proto of compilation cache 2021-11-27 16:02:01 +08:00
liuyang_655 1865d98689 final ckpt 2021-11-27 02:56:48 -05:00
i-robot c1798df274 !26837 tag environment bugfix
Merge pull request !26837 from chenweifeng/tag-environment-bug-fix
2021-11-27 07:44:27 +00:00
i-robot 01b1a53192 !26862 add adapter of HSwish and HSwishGrad for graphengine.
Merge pull request !26862 from wangshuide/wsd_master_adapter
2021-11-27 07:41:33 +00:00
zhaosida 7de4c114b3 fix avx512 compile option 2021-11-27 15:40:35 +08:00
i-robot f85dea2959 !26662 fix lu kernel
Merge pull request !26662 from zhuzhongrui/gmres
2021-11-27 07:18:46 +00:00
huanghui c4d61629b5 fix wrong scope and debug info 2021-11-27 14:50:16 +08:00
i-robot 6213135cd8 !26872 modify lossmonitor and print convert type
Merge pull request !26872 from changzherui/mod_lossmonitor_and_print
2021-11-27 06:39:39 +00:00
i-robot 2bd6deaeeb !26682 GraphKernel For CPU
Merge pull request !26682 from ZengZitao/gk_cpu
2021-11-27 06:23:01 +00:00
i-robot e2b5ee71c2 !26865 Refactor original kernal name so all output tensors are included in stat dump
Merge pull request !26865 from Jimmy Qi/stat-hotfix
2021-11-27 04:22:42 +00:00
dayschan 4b19d483a3 move graph_kernel_cluster to directory graph_kernel/core 2021-11-27 11:55:58 +08:00
i-robot 4a82477626 !26794 transformer_slice_activation_config_fix
Merge pull request !26794 from yao_yf/add_transformer_slice_activation_config_fix
2021-11-27 03:42:36 +00:00
liangzelang 7255af4e07 No need to handle control flow except multigraph sink. 2021-11-27 11:42:04 +08:00
changzherui 4260b5e5d9 modify lossmonitor and print convert type 2021-11-27 11:41:03 +08:00
caifubi c32e63bfc8 output without monad 2021-11-27 11:39:43 +08:00
vddong 91d3a7543f [feat][assistant][I3T96F] add new data operator SpeechCommands 2021-11-27 11:25:18 +08:00
i-robot 6f95837c0e !26825 [Fallback] supports the use of attr/method on the interpreted nodes
Merge pull request !26825 from huangbingjian/support_numpy_method
2021-11-27 03:07:24 +00:00
i-robot cb7847c9dd !26819 add jit config param for compile and fix summary graph
Merge pull request !26819 from jiangshuqiang/fix_summary_graph_part2
2021-11-27 02:19:43 +00:00
i-robot b1deeb425d !26849 bind stream with handle
Merge pull request !26849 from zhujingxuan/master
2021-11-27 02:11:40 +00:00
He Wei 27981a28f2 Fix map::overflow when save_graphs enabled
AnfExporter use a user defined hasher in its param-to-index map,
this will lead robin-hood hash map overflow. We fix this by change
param-to-index map to use a OrderedMap with std::unordered_map inside.
2021-11-27 10:06:58 +08:00
i-robot 8c5957c387 !26831 fix output tensor num bug
Merge pull request !26831 from caifubi/master-pynative-lazy-build-bug
2021-11-27 01:15:56 +00:00
i-robot 4252b24335 !26792 malloc ts memory for label
Merge pull request !26792 from zhoufeng/change-label-memory-type
2021-11-27 01:09:03 +00:00
i-robot 9c9ccdff84 !26823 Fix FusionOp nodes null value issue in offline debugger
Merge pull request !26823 from TinaMengtingZhang/bugfix_fusionop
2021-11-27 01:06:35 +00:00
i-robot 0ce94030e8 !26859 update api format 1126
Merge pull request !26859 from yingchen/code_docs_api1126
2021-11-27 01:05:59 +00:00
i-robot 16a2535205 !26820 optimizes the kernel error description of LSTM, Pad, ReLU, etc.
Merge pull request !26820 from wangshuide/wsd_master_new
2021-11-27 00:53:58 +00:00
John Tzanakakis 64893bf9eb deal with kernel_graph_#_ prefix in tensor name in overflow wp checking 2021-11-26 18:51:20 -05:00
Jimmy Qi 2a4a98ce9d Refactor original kernal name so all output tensors are included in stat
dump
2021-11-26 19:21:54 +00:00
i-robot 0b236f91b9 !26822 Ignore inline ops in Autotune
Merge pull request !26822 from h.farahat/fix_inlineops_AT
2021-11-26 15:12:42 +00:00
wangshuide2020 9a057fce14 add adapter of HSwish and HSwishGrad for graphengine. 2021-11-26 20:18:45 +08:00
jiangshuqiang c9a0edd24e add jit config param for compile and fix summary graph 2021-11-26 04:04:05 -08:00
yingchen 4d46962b11 update api format 2021-11-26 19:23:40 +08:00
zengzitao 62458b5636 adapt graph kernel for cpu 2021-11-26 18:42:34 +08:00
zhujingxuan 30c6fa7f9b bind stream with handle 2021-11-26 17:44:16 +08:00
i-robot d1e4e674ab !26690 Add supoort resolving outer lambda function for ops.Partial.
Merge pull request !26690 from hezhenhao1/add_lambda
2021-11-26 09:21:34 +00:00
wangshuide2020 6cbe8dd02e optimizes the kernel error description of LSTM, Pad, ReLU, etc. 2021-11-26 16:49:20 +08:00
huangbingjian cf4a4ba900 Partially support numpy attr/method 2021-11-26 16:46:55 +08:00
Bingya Weng ce85cba866 add chinese api of probability 2021-11-26 16:42:29 +08:00
xiefangqi 0570664dfa modify threshold of batch memory usage to 0.8 2021-11-26 16:35:12 +08:00
wanyiming 822b11a9be code_docs_cnapi 2021-11-26 16:30:37 +08:00
z00512249 872d8b37da fix lu factor and solve for cpu backend 2021-11-26 16:24:01 +08:00
i-robot 6ab8ddfd3a !26806 Modify updateModel for cross-silo STABLE_PW_ENCRYPT
Merge pull request !26806 from jxlang910/modify_updateModel
2021-11-26 08:12:38 +00:00
i-robot c1b1324ca4 !26815 310 ocr model support dynamic hw
Merge pull request !26815 from zhengyuanhua/model_br1
2021-11-26 08:00:09 +00:00
Zichun Ye 8398c07d68 update random normal op impl to speed up sampling 2021-11-26 15:43:36 +08:00