提交 · 1f0436967c9113ec202b67c349dd72a8adf6ddd5 · sadsadfsdsfdsfd / MegEngine

27 7月, 2021 4 次提交
- M
  refactor(mge/distributed): using nccl as default in distributed training · 1f043696
  由 Megvii Engine Team 提交于 7月 19, 2021
```
GitOrigin-RevId: 81268e84bc93e060b9436b4b940b7286e353660f
```
  1f043696
- M
  feat(mge/distributed): deprecate get_device_count_by_fork · b17a02d4
  由 Megvii Engine Team 提交于 7月 19, 2021
```
GitOrigin-RevId: 6147c3ae905c65a532fe000aa51546abdfddc96f
```
  b17a02d4
- M
  build(dnn/cutlass): fix build for cutlass · f8b0f2cb
  由 Megvii Engine Team 提交于 7月 19, 2021
```
GitOrigin-RevId: 9aa095fe84e7f01646280b49bc268e52e4fb6909
```
  f8b0f2cb
- K
  
  fix(ci): git set user and email · 0fb4e9a9
  由 konghuanjun 提交于 6月 24, 2021
  
  0fb4e9a9
23 7月, 2021 6 次提交
- H
  
  feat(mge/third_party): update MegRay version · 6af4a32e
  由 huangxinda 提交于 7月 19, 2021
  
  6af4a32e
- H
  
  feat(mge/third_party): update cutlass version · 093f7ae7
  由 huangxinda 提交于 7月 19, 2021
  
  093f7ae7
- M
  chore(release): bump version · c2daea3c
  由 Megvii Engine Team 提交于 7月 18, 2021
```
GitOrigin-RevId: a016ea9d564a32adecaf6a80bc01a8ddb95d88db
```
  c2daea3c
- M
  chore(mge): run get_device_count("gpu") in subprocess · 207a3463
  由 Megvii Engine Team 提交于 1月 19, 2021
```
GitOrigin-RevId: 0f0dc001cfc45fc0d04de1a86c27f8bba8185d6b
```
  207a3463
- M
  perf(mgb): disable FoldingConvBiasDimshufflePass in cuda10 for performance · 869a0327
  由 Megvii Engine Team 提交于 7月 13, 2021
```
GitOrigin-RevId: d1b95a6f01ba73f98c0094e00fee3e61e9139628
```
  869a0327
- L
  Merge pull request #175 from tpoisonooo:fix-spell-error · 0baf6b0d
  由 liuke 提交于 7月 23, 2021
```
GitOrigin-RevId: fe649676ac1aa844a469d1cbf19dd7612cc0f29f
```
  0baf6b0d
19 7月, 2021 30 次提交
- M
  fix(mgb/gopt): fix testcase for enable nchw64 pass · 239916a9
  由 Megvii Engine Team 提交于 6月 16, 2021
```
GitOrigin-RevId: 2ae8d1608d7e9623f4c2abe59f9c80d54559aaf4
```
  239916a9
- M
  feat(mgb/gopt): support nhwc conv in tensor reformat pass · 2ab5c53f
  由 Megvii Engine Team 提交于 6月 09, 2021
```
GitOrigin-RevId: 43e78d758ab352c9e47d9ca1bb5fe868d4443458
```
  2ab5c53f
- M
  feat(mgb/gopt): modify padding policy for 4bit conv bias oprs · 009c90a2
  由 Megvii Engine Team 提交于 6月 08, 2021
```
GitOrigin-RevId: 188a2c3728c017c77eba433211b308c9680b1dad
```
  009c90a2
- M
  feat(dnn/cuda): generate cutlass kimpls using cmake and bazel · 4eda3388
  由 Megvii Engine Team 提交于 6月 21, 2021
```
GitOrigin-RevId: da3bcfb85af3f148116b8b6a51f214c6cef8684e
```
  4eda3388
- M
  fix(dnn/cuda): fix testcase for fallback nchw qs8 conv · 8d248a6a
  由 Megvii Engine Team 提交于 6月 23, 2021
```
GitOrigin-RevId: 646440db59f0157a3fdbd8061167f9ac04dbd422
```
  8d248a6a
- M
  feat(dnn/cuda): add relayout format kernel for nchw <-> nhwc · 894a2407
  由 Megvii Engine Team 提交于 6月 04, 2021
```
GitOrigin-RevId: e11f3e54085929ab9919fe2070fcc9e633755b69
```
  894a2407
- M
  refactor(dnn/cuda): refactor relayout format kernels · 43c59204
  由 Megvii Engine Team 提交于 5月 31, 2021
```
GitOrigin-RevId: ab86e6653342ae9f74dd069cfb85aabca6dc637c
```
  43c59204
- M
  feat(dnn/cuda): add nhwc int4 conv support · f41a8086
  由 Megvii Engine Team 提交于 6月 11, 2021
```
GitOrigin-RevId: 5236b235d0310feecac0fbd1dc76ff0755cf9426
```
  f41a8086
- M
  refactor(dnn/cuda): refactor cutlass kernel generator for gemm and gemv · 5a14a892
  由 Megvii Engine Team 提交于 5月 31, 2021
```
GitOrigin-RevId: 11d78ab2270f0720d7d79e186124a1254c467980
```
  5a14a892
- M
  refactor(dnn/cuda): refactor cutlass kernel generator for deconv operation · b33217d8
  由 Megvii Engine Team 提交于 5月 27, 2021
```
GitOrigin-RevId: 88e962a912dc0f31b8b31787a9ab7aa1b38d1264
```
  b33217d8
- M
  refactor(dnn/cuda): refactor kernel generator for cutlass convolution kernels · 4abf7bd3
  由 Megvii Engine Team 提交于 5月 27, 2021
```
GitOrigin-RevId: 7882f9c68c0e0f35159f12a07ccdf027f4540ff1
```
  4abf7bd3
- M
  feat(dnn/cuda): add convolution with i8 input and u4 output · b4687ce8
  由 Megvii Engine Team 提交于 5月 24, 2021
```
GitOrigin-RevId: 8be439abf1f448a6be33ea0e57c48e674c5c94c4
```
  b4687ce8
- M
  fix(dnn/cuda): fix recursive algo search for fallback_nchw_qs8 · 00083d13
  由 Megvii Engine Team 提交于 5月 20, 2021
```
GitOrigin-RevId: 6be2991224bced3a38a17b6b888fd4f324d03f9f
```
  00083d13
- M
  feat(mgb/gopt): add fusion support for conv, astype(s4) and reformat · bba04f02
  由 Megvii Engine Team 提交于 5月 18, 2021
```
GitOrigin-RevId: 6329ca2c5fe85dc4f7c94610a293ab506a3c2945
```
  bba04f02
- M
  feat(dnn/cuda): add convolution with i8 input and i4 output · 66f70578
  由 Megvii Engine Team 提交于 5月 18, 2021
```
GitOrigin-RevId: 10512645d5d5ac3d985720788760bf8a3855c1f1
```
  66f70578
- M
  feat(gopt/inference): allow Float32 output dtype in EnableNCHW64Pass · 6d686ff2
  由 Megvii Engine Team 提交于 6月 16, 2021
```
GitOrigin-RevId: 1891efb76f66a6abbd0a56820281b4fe91e70304
```
  6d686ff2
- M
  feat(gopt/inference): allow Float32 output dtype in EnableNCHW4Pass · 7d3df995
  由 Megvii Engine Team 提交于 6月 11, 2021
```
GitOrigin-RevId: 81100dbaf7d9e8f5aade6c630ead79ea5ba45e52
```
  7d3df995
- M
  fix(dnn/cuda): fix AlgoFallbackNCHWQS8 to support Float32 dst · 633016a9
  由 Megvii Engine Team 提交于 6月 10, 2021
```
GitOrigin-RevId: 06f90f5cf384bc4ddb2f97860e4f530ee9a85705
```
  633016a9
- M
  feat(opr): add bn backward for inference mode · e6caa9ff
  由 Megvii Engine Team 提交于 6月 01, 2021
```
GitOrigin-RevId: bb643cb62fbba90ca8846a3550f88bf6763ddd58
```
  e6caa9ff
- X
  
  test(mge): delete test_external.py · c90fa087
  由 Xinda Huang 提交于 7月 17, 2021
  
  c90fa087
- M
  fix(imperative/module): remove ``__getattribute__`` method in module · b2944559
  由 Megvii Engine Team 提交于 7月 03, 2021
```
GitOrigin-RevId: 5ac525f010eb0d23d3f83788709fe1360f03e9f3
```
  b2944559
- M
  fix(src/serialization): fix compatibility error of oss model · 77ead937
  由 Megvii Engine Team 提交于 7月 13, 2021
```
GitOrigin-RevId: 43e0fa4fe14cbfdcbc835a35487368098e9772ed
```
  77ead937
- M
  fix(imperative): remove convert_inputs · 070c8117
  由 Megvii Engine Team 提交于 7月 07, 2021
```
GitOrigin-RevId: a3c43db746789bb9ba418c91578c16368a96536c
```
  070c8117
- M
  docs(mge): refactor docs to remove warnings · f40df602
  由 Megvii Engine Team 提交于 7月 09, 2021
```
GitOrigin-RevId: efefc2a4a21027600d1ac3156c648bb987d039e4
```
  f40df602
- M
  fix(mge/functional): fix F.topk(kth_only=True) · 1040b778
  由 Megvii Engine Team 提交于 6月 28, 2021
```
GitOrigin-RevId: ddecd1d14b62b43b2934fa2671562b04495ff5b4
```
  1040b778
- M
  docs(distributed.functional): add return type for all_reduce_max (jira #MGE-2706) · 551cc701
  由 Megvii Engine Team 提交于 7月 13, 2021
```
GitOrigin-RevId: a29f1f1880674a45fa593582b34b807be1e46786
```
  551cc701
- M
  feat(docs): add docs for megengine.functional.ones_like(jira #MGE-2702) · 72ff7aec
  由 Megvii Engine Team 提交于 7月 13, 2021
```
GitOrigin-RevId: e808fc9e4e0f4614582e5767000545ea87738734
```
  72ff7aec
- M
  fix(mge/random): fix random seed · 7c9569e4
  由 Megvii Engine Team 提交于 7月 09, 2021
```
GitOrigin-RevId: 121f459b1b30a086776e9832e15849de9987da0e
```
  7c9569e4
- M
  fix(mgb): remove static mem record from tee · 07de1571
  由 Megvii Engine Team 提交于 7月 06, 2021
```
GitOrigin-RevId: ac61b2a5ebb813ed183048a5b5d88dc56bd589d6
```
  07de1571
- M
  test(mge/fakequant): use fixed input for lsq test to temperarily avoid precision error · d7b6bfd5
  由 Megvii Engine Team 提交于 7月 14, 2021
```
GitOrigin-RevId: e91c71874eb16e8535497ec7c5840d08cf1a3d2b
```
  d7b6bfd5

sadsadfsdsfdsfd / MegEngine 与 Fork 源项目一致

sadsadfsdsfdsfd / MegEngine
与 Fork 源项目一致