7083a00b34
(trunk/babel/s5b) Removing one more old BNF script
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3656 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-03 00:38:09 +00:00
fbc7c13d55
(trunk/babel/s5b) Adding a new semi-supervised training script and removing old BNF scripts
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3655 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-03 00:37:09 +00:00
5c63969966
(trunk/babel/s5b) Lots of fixes around the files. Removed the old script for semisupervised training, as we already have a new one
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3653 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-02 22:36:48 +00:00
1bc6f3821d
trunk/egs/babel/s5b: Modified conf file for semi-supervised training of DNN
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3652 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-02 16:26:59 +00:00
c1ace00d33
trunk/egs/babel/s5b: Minor modifications of commands in script run-6-semisupervised-seg.sh
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3651 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-02 15:47:16 +00:00
3ec5bf91ca
(trunk/babel/s5b) Changing configs to contain link to the RADICAL UEM segmentation
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3648 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-02 03:28:18 +00:00
dae25f7041
trunk: BABEL script: removing some old, unused scripts.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3647 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-02 00:09:21 +00:00
390127bbb3
Small fix to proxy keywords generation to avoid blowup.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3643 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-01 05:17:34 +00:00
4b2173c845
(trunk/babel/s5b) Fixes in the Hatian eval descriptions
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3640 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-03-01 00:01:36 +00:00
6b75ec4806
(trunk/babel/s5b) Adding configs for eval and shadow datasets.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3637 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 22:07:07 +00:00
bf746bd057
fixing chime run.sh after nnet scripts moved
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3636 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 21:47:32 +00:00
004b22be57
fixing aurora run.sh after nnet scripts moved
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3635 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 21:38:30 +00:00
0a2102e2b9
trunk,nnet: removing svn:mime_type property
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3631 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 17:58:59 +00:00
2c50f1d052
Adding minimization to phone determinization, which will speed up the downstream operations on the determinized lattices.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3630 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 17:57:30 +00:00
50517c2f19
trunk,nnet: WARNING! MOVING KARELS NN SCRIPTS! THIS CAN BREAK YOUR CUSTOM RECIPES!
...
* moving scripts related to Karel's DNN recipe to from steps/* to steps/nnet/*,
* updating the DNN recipes (rm,wsj,swbd), new features:
** can be resumed by --stage,
** renaming experiments to exp/dnn*,
** using scripts at its new location,
We are syncing the script directory structure to be same as in Dan's recipe,
if you encounter any difficulties, please let me know.
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3629 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 17:56:03 +00:00
c7443c8f26
trunk: Minor fix in script egs/babel/s5b/local/combine_posteriors.sh to handle single decode directory. Minor fix in initialization of vectors in src/bin/vector-sum.cc.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3628 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-28 17:16:50 +00:00
c7a26d90e4
trunk/egs/babel: Adding a script steps/nnet2/update_nnet.sh that updates an existing nnet model without initializing it. Minor fix added in script run-6-semisupervised-seg.sh to use update_nnet.sh
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3616 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-27 21:22:40 +00:00
417778ab25
trunk/egs/babel/steps/nnet2: Modified get_lda.sh and train_pnorm.sh to optionally use FMLLR transforms from a directory different from the alignment directory
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3614 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-27 20:28:07 +00:00
6701a5ca18
trunk/egs: Small fix to steps/make_denlats_nnet.sh (thanks to Feiteng Li)
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3613 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-27 06:10:56 +00:00
1b53a9c232
trunk:egs readme update on the removal of s5b
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3612 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-27 03:46:35 +00:00
19107f6556
trunk:src remove since not all data are released
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3611 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-27 03:34:17 +00:00
f55e8d0f42
fix '-ldl' error specific to gcc-4.6.3, ubuntu 12.04 by changing fst rule in tools/Makefile
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3609 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-26 23:10:11 +00:00
a9396adaf6
trunk: remove limitation on split not-shared tree building; remove some stuff that was never finished RE AWS; minor changes to run.sh, cosmetic.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3600 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-26 03:06:42 +00:00
ca07e7b7ee
Added determinization after KxL2xE for proxy keywords; helps the speed, but not fast enough for the 1million lexicon
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3594 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-25 17:48:59 +00:00
725f4abd68
WARNING: this change list changed GauPost to GaussPost, and also made it pdf-id indexed instead of transition-id indexed. Details: 1) Modified those posterior related programs that use TransitionIdToPdf to get a pdf-id, and later on only use the pdf-id. We merge the posteriors that corresponds to the same pdf-id to avoid redundant computation. 2) Modified phone lattice determinization, added a wrapper for the lattice type determinization to reduce redundant code in the decoding binaries.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3588 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-25 04:00:06 +00:00
cd36df98b5
trunk:src remove unused file
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3587 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-24 22:41:27 +00:00
24310794f9
trunk: minor change to some functions manipulating posteriors, for efficiency; efficiency improvements in apply-cmvn-sliding; cosmetic changes.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3585 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-24 22:16:33 +00:00
0db68a27f5
trunk/egs/babel: Bug fix in script for getting examples for semi-supervised training of DNN - get_egs_semi_supervised.sh
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3579 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-24 16:48:39 +00:00
0103c446dc
trunk/egs/voxforge/s5: Small changes to improve acoustic model training.
...
Fixes the issues outlined in:
http://sourceforge.net/p/kaldi/discussion/1355348/thread/4db46866/#941f/4f9c/cf07
The test time LM is now built on all available transcripts, including those for
the test utterances, which is not the correct way to evaluate a system, but
for this particular recipe I don't think it matters much.
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3572 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-23 12:19:49 +00:00
7ea3dff539
trunk:src not ready, nnet/nnet-cache-tgtmat.h not exist
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3571 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-23 10:15:28 +00:00
a2a1b3d0c3
trunk: renaming --zero-if-disjoint option to --drop-frames in calling scripts (should have been done a long time ago, when I renamed the option)
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3567 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-21 19:52:26 +00:00
c191a76f82
trunk/egs/babel: Minor fix in segmentation script generate_segments.sh
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3566 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-21 18:13:33 +00:00
6a2db5e1b7
trunk/egs/babel: Minor change in segmentation.py to take python location from the environment
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3565 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-21 16:41:43 +00:00
6d45c140a6
trunk: disabling some tests, which are not ready yet
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3564 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-21 12:39:36 +00:00
b42b9b733f
trunk: BABEL: changing default acoustic scale for decoding bottleneck system (thanks to Pegah Ghahremani); minor fix in train_deltas.sh (thanks to Simon Kluepfel)
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3562 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-21 04:19:14 +00:00
efb56ad7d7
trunk: Minor, mostly cosmetic fixes
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3559 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-21 03:47:30 +00:00
22285c7d2c
trunk/egs/babel: Minor change in segmentation script generate_segments.sh
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3557 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 23:18:52 +00:00
ebdb816f84
(trunk/babel/s5b) Adding unsup dataset (dataset for unsupervised/semisupervised training)
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3556 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 22:35:14 +00:00
6f388b925d
trunk: minor fix to scripts train_pnorm.sh and train_pnorn_ensemble.sh to work with non-GMM input model
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3553 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 18:07:31 +00:00
9692cf08a5
(trunk/babel/s5b) Fix the bug accidentaly introduced when renaming the bnf directories (removing the _app) suffix
...
Because the param was stored in the plp directory and the way the param names
are constructed, this lead to overwriting the PLP parametrization of the same
dataset. Now the BNF param is stored in the directory param_bnf (plp_bnf seemed
silly)
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3552 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 17:51:10 +00:00
a6688e9daa
trunk,nnet: updating the codebase
...
- removing 'Cache' class, replaced by 'Randomizer'
- removing obsolete training tools using 'Cache' class
- adding 2d convolutional components from Harish
- removing scaled inv-prior-weighted cross-entropy Class, not helpful
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3551 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 10:55:18 +00:00
3524adaac2
(trunk)
...
Updating semi-supervised DNN training scripts and associated programs in trunk.
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3550 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 07:28:08 +00:00
7318249685
fixed a typo in run-2a-nnet-ensemble.sh
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3549 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 06:44:10 +00:00
d1eb28997b
(trunk/babel/s4b) Added support for data sources specified using multiple locations and filelists
...
This tremendously simplifies handling of untranscribed data and the shadow set processing
In the longer term, we will be able to remove the shadow-set specific code paths in decoding.
For example, how to specify the data source using multple locations, see the example config
that is part of this commit
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3546 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-20 06:07:12 +00:00
717ad1c436
trunk: Change name of program sum-vecs to vector-sum (was not used, but is now needed; new name is more discoverable). Minor changes in usage messages
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3541 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-19 00:08:45 +00:00
d081efc2be
(trunk)
...
Adding scripts and source files for semi-supervised DNN training to trunk
The scripts affect only the Babel recipe.
Adds additional programs in src/bin and src/nnet2bin to get examples
for semi-supervised training.
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3540 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-18 23:58:34 +00:00
f9c8bdc451
(trunk/babel/s5b) Additional fixes in the new decoding pipeline +improvements for the combination scripts.
...
The kws_combine.sh now supports supplying individual weights of the systems (before this change, all the systems
were taken with the same weight). The score_combine.sh now properly filters <hes> tag (which is marked optionable
deletable in the eval documentation.
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3538 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-18 21:06:20 +00:00
228a61d486
(trunk/babel/s5b) Mostly fixes to the new decoding script. New feature: added OOV search.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3537 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-18 20:54:57 +00:00
7ea7b9f2c7
trunk: Script fix in train_quick.sh (thanks to Simon Klupfel); and clarifications about how to turn off CMN in scripts.
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3536 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-18 19:40:09 +00:00
1e2af4f0ef
Segmentation scripts
...
git-svn-id: https://svn.code.sf.net/p/kaldi/code/trunk@3535 5e6a8d80-dfce-4ca6-a32a-6e07a63d50c8
2014-02-18 03:40:48 +00:00
Jan Trmal
Vimal Manohar
Dan Povey
Guoguo Chen
Chao Weng
Karel Vesely
Vassil Panayotov
Ho Yin Chan
Ondrej Platek
Xiaohui Zhang