Commit Graph

44 Commits

Author SHA1 Message Date
Tomasz Dwojak
b6c85fdfb4 Change README to markdown 2016-04-18 22:47:41 +02:00
Tomasz Dwojak
49233da2bf Update README 2016-04-18 22:36:17 +02:00
Tomasz Dwojak
9d7e486e2a Clean mblas implementation - remove uneeded files 2016-04-18 21:53:52 +02:00
Marcin Junczys-Dowmunt
f2d298e9a6 got rid of unneeded encapsulation 2016-04-18 21:29:48 +02:00
Marcin Junczys-Dowmunt
b43747ab5c organize lm queries in to batch 2016-04-18 21:16:07 +02:00
Ubuntu
349a0266cf removed cudaDevicesynchronize 2016-04-18 18:44:22 +00:00
Ubuntu
890e3fe066 Merge branch 'master' of https://github.com/emjotde/amunn 2016-04-18 18:18:35 +00:00
Ubuntu
74c1827c21 rearrange model loading 2016-04-18 18:18:24 +00:00
Marcin Junczys-Dowmunt
d7b00d22e7 enable lm theads 2016-04-18 20:17:57 +02:00
Marcin Junczys-Dowmunt
1b36ce7e89 missing index 2016-04-18 19:57:01 +02:00
Marcin Junczys-Dowmunt
84bad30b20 merge 2016-04-18 19:55:41 +02:00
Marcin Junczys-Dowmunt
1722f4cbb0 synchronize message 2016-04-18 19:54:46 +02:00
Kenneth Heafield
30451239b8 Shebang 2016-04-18 14:55:25 +00:00
Kenneth Heafield
89bc8b1d36 User-specified kenlm path 2016-04-18 14:49:35 +00:00
Marcin Junczys-Dowmunt
822b1e1d24 multi-gpu works with multiple threads but kenlm breaks 2016-04-16 20:13:03 +02:00
Marcin Junczys-Dowmunt
c4d9466727 Merge branch 'master' of github.com:emjotde/amunn 2016-04-16 16:44:08 +02:00
Marcin Junczys-Dowmunt
3264a8ab6f lots of stuff, mostly chaotic, kenlm sort of works 2016-04-16 16:44:03 +02:00
Tomasz Dwojak
7c69e2500f Speed up bpe 2016-04-16 10:11:23 +02:00
Tomasz Dwojak
f2644cd61d Optimalize bpe 2016-04-16 00:12:18 +02:00
Tomasz Dwojak
208fb0565f Add utf-8 support to bpe 2016-04-15 22:50:12 +02:00
Tomasz Dwojak
44ea08236f Add bpe app 2016-04-15 21:34:28 +02:00
Tomasz Dwojak
23bff0a14c Add working bpe 2016-04-15 21:18:33 +02:00
Tomasz Dwojak
62e074e081 Add python's pyc files to ignore 2016-04-15 16:27:48 +02:00
Tomasz Dwojak
ec5f9d9080 Add an initla verison od bpe 2016-04-15 16:27:47 +02:00
Marcin Junczys-Dowmunt
8b9bb9ded3 new cmake version finally works 2016-04-15 15:20:31 +02:00
Marcin Junczys-Dowmunt
c7c2165bda use max_element for smaller beam sizes 2016-04-15 01:43:01 +02:00
Marcin Junczys-Dowmunt
0355760f1b Merge branch 'master' of github.com:emjotde/amunn 2016-04-15 01:03:57 +02:00
Marcin Junczys-Dowmunt
4e4ee95633 clean-up, get rid of GH models 2016-04-15 01:03:49 +02:00
Marcin Junczys-Dowmunt
9c2c494c14 Update README 2016-04-15 00:01:31 +01:00
Marcin Junczys-Dowmunt
1da629da20 added missing file 2016-04-15 01:00:20 +02:00
Marcin Junczys-Dowmunt
d7d8b44bbb working ensemble and n-best-list score breakdown 2016-04-15 00:27:34 +02:00
Marcin Junczys-Dowmunt
136aaea528 working model ensemble 2016-04-14 23:06:53 +02:00
Marcin Junczys-Dowmunt
1505d3811f finished n-best lits 2016-04-14 21:27:15 +02:00
Marcin Junczys-Dowmunt
f156da98e9 types.h 2016-04-14 20:52:11 +02:00
Marcin Junczys-Dowmunt
19e71f8061 towards n-best-lists 2016-04-14 20:49:38 +02:00
Marcin Junczys-Dowmunt
436f0bd52e note about vocabularies 2016-04-14 17:25:05 +02:00
Marcin Junczys-Dowmunt
67c620a7a4 Readme 2016-04-14 17:21:07 +02:00
Marcin Junczys-Dowmunt
d5cb8578f2 work around for CUDA 7.0 bug 2016-04-14 15:05:05 +02:00
Marcin Junczys-Dowmunt
93a1590035 work around for CUDA 7.0 bug 2016-04-14 15:02:08 +02:00
Marcin Junczys-Dowmunt
e3f94e10ec remove boost::split 2016-04-14 14:55:14 +02:00
Marcin Junczys-Dowmunt
3bc0485b73 remove boost::split 2016-04-14 14:53:40 +02:00
Marcin Junczys-Dowmunt
14457d0c58 turn into objects 2016-04-14 14:29:41 +02:00
Marcin Junczys-Dowmunt
945e7c8a43 first commit 2016-04-14 13:44:55 +02:00
Marcin Junczys-Dowmunt
edec54a6d0 Initial commit 2016-04-14 11:27:41 +01:00