Commits · 4641030e9d0cf682b09a2dad4864dcfdcc1d6534 · Emmay / NiuTrans.Tensor · GitLab

18 Mar, 2019 2 commits
- add headers in cpp and fix the bug of MAX_LEN in XLink.h · 4641030e
  Tianzhi committed Mar 18, 2019
  
  4641030e Browse Files
- use OpenBLAS API to accelerate the speed of CPU · ba8bc234
  Tianzhi committed Mar 18, 2019
  
  ba8bc234 Browse Files
14 Mar, 2019 1 commit
- class of predictors · b9871b8d
  xiaotong committed Mar 14, 2019
  
  b9871b8d Browse Files
13 Mar, 2019 3 commits
- clean the code · 8cb65ef5
  xiaotong committed Mar 13, 2019
  
  8cb65ef5 Browse Files
- create source files of search module · d2011ab6
  xiaotong committed Mar 13, 2019
  
  d2011ab6 Browse Files
- implement MulAndShift and bug fixs · 3bf085db
  姜雨帆 committed Mar 13, 2019
  
  3bf085db Browse Files
22 Jan, 2019 1 commit
- clean the code · 411faffa
  xiaotong committed Jan 22, 2019
  
  411faffa Browse Files
06 Jan, 2019 1 commit
- sort samples by max length of source and target sequences · f04dfc79
  xiaotong committed Jan 06, 2019
  
  f04dfc79 Browse Files
31 Dec, 2018 1 commit
- distribute samples into buckets wrt length · 12d11fab
  xiaotong committed Dec 31, 2018
  
  12d11fab Browse Files
30 Dec, 2018 2 commits
- bug fix · c260abfb
  xiaotong committed Dec 30, 2018
  
  c260abfb Browse Files
- randomize sample batches rarther than loading them is sorted manner · 2161f65b
  xiaotong committed Dec 30, 2018
  
  2161f65b Browse Files
29 Dec, 2018 3 commits
- dropout withno use of broadcasting · 52a50ab2
  xiaotong committed Dec 29, 2018
  
  52a50ab2 Browse Files
- improve dropout implementation by broadcasting from different dimensions · b1a9adde
  xiaotong committed Dec 29, 2018
  
  b1a9adde Browse Files
- random number fixes · 896e5231
  xiaotong committed Dec 29, 2018
  
  896e5231 Browse Files
28 Dec, 2018 3 commits
- better implementation of dropout · 52a27964
  xiaotong committed Dec 28, 2018
  
  52a27964 Browse Files
- broadcasting · a8304bed
  xiaotong committed Dec 28, 2018
  
  a8304bed Browse Files
- show memory usage · 14f245fa
  xiaotong committed Dec 28, 2018
  
  14f245fa Browse Files
27 Dec, 2018 2 commits
- better definition of "batching" · 1739a22f
  xiaotong committed Dec 27, 2018
  
  1739a22f Browse Files
- the first version of _SumBroadcast · 32e78bda
  xiaotong committed Dec 27, 2018
  
  32e78bda Browse Files
26 Dec, 2018 2 commits
- implement sum broadcasting by unsqueezing · f784b909
  xiaotong committed Dec 26, 2018
  
  f784b909 Browse Files
- add broadcasting · 6ba61279
  xiaotong committed Dec 26, 2018
  
  6ba61279 Browse Files
25 Dec, 2018 1 commit
- remove -smallfootprint and keep the net as small as possible by default · 947db8fb
  xiaotong committed Dec 25, 2018
  
  947db8fb Browse Files
24 Dec, 2018 2 commits
- 1. fixed the bug in the Spread function. 2. add compare and OnehotAndIndex… · 854a4835
```
1. fixed the bug in the Spread function.  2. add compare and OnehotAndIndex function 3. optimize the code
```
  xuchen committed Dec 24, 2018
  854a4835 Browse Files
- introduce small-footprint backward computation (by using "-smallfootprint") · 61c4d15c
  xiaotong committed Dec 24, 2018
  
  61c4d15c Browse Files
18 Dec, 2018 1 commit
- replace XLink::ShowNetwork with XNet::SHowNetwork · f8c57160
  xiaotong committed Dec 18, 2018
  
  f8c57160 Browse Files
13 Dec, 2018 1 commit
- add a start position for setdata · 31d989e9
  xiaotong committed Dec 13, 2018
  
  31d989e9 Browse Files
09 Dec, 2018 1 commit
- generate masks for t2t mt models · 3e7f7645
  xiaotong committed Dec 09, 2018
  
  3e7f7645 Browse Files
07 Dec, 2018 2 commits
- softmax output -> logsoftmax output · 21892dbf
  xiaotong committed Dec 07, 2018
  
  21892dbf Browse Files
- rewrite reducesumall · 8f2d16f5
  xiaotong committed Dec 07, 2018
  
  8f2d16f5 Browse Files
04 Dec, 2018 1 commit
- fix the bug of suming the word numbers of both source and target-sides for the… · 9b0d4df8
```
fix the bug of suming the word numbers of both source and target-sides for the loss computation in transformer MT models
```
  xiaotong committed Dec 04, 2018
  9b0d4df8 Browse Files
03 Dec, 2018 1 commit
- better control of the strategy of the merge method · 1c26ff5b
  xiaotong committed Dec 03, 2018
  
  1c26ff5b Browse Files
29 Nov, 2018 1 commit
- revise the comment · 78954fad
  xiaotong committed Nov 29, 2018
  
  78954fad Browse Files
28 Nov, 2018 8 commits
- clean useless code · dc56fd67
  xiaotong committed Nov 28, 2018
  
  dc56fd67 Browse Files
- recover the Main.cpp file · 49869ae5
  xuchen committed Nov 28, 2018
  
  49869ae5 Browse Files
- optimize the implementation of LoadBatchMT · 126c581f
  xuchen committed Nov 28, 2018
  
  126c581f Browse Files
- cpu implementation of the gather and spread function · b3fa7a37
  xuchen committed Nov 28, 2018
  
  b3fa7a37 Browse Files
- improve the implementation of gather and spread · 2fea6615
  xuchen committed Nov 28, 2018
  
  2fea6615 Browse Files
- replace the cuda thread allocation code · 99225c29
  xiaotong committed Nov 28, 2018
  
  99225c29 Browse Files
- remove commented lines · 58abfba4
  xiaotong committed Nov 28, 2018
  
  58abfba4 Browse Files
- better memory allocation · 4cabf059
  xiaotong committed Nov 28, 2018
  
  4cabf059 Browse Files