generic fst supports constfst and vectorfst (without FstHolder) #1661
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
Contributor
| Thanks! I'll check it when you say the mini-librispeech tests are done. …On Wed, May 31, 2017 at 8:20 PM, LvHang ***@***.***> wrote: This PR goes on #1548 <#1548> This version has removed the FstHolder. That means only "if" branch in *-latgen-faster is fixed in order to speeding up and saving memory about decoding. @danpovey <https://github.com/danpovey> The test about nnet3 tdnn training has passed with swbd corpus. Now, I'm testing the whole pipeline(run.sh) with mini_librispeech corpus. I will notice you when it accomplish. Bests, Hang ------------------------------ You can view, comment on, or merge this pull request online at: #1661 Commit Summary - generic fst supports constfst and vectorfst (no FstHolder) File Changes - *M* egs/wsj/s5/utils/mkgraph.sh <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-0> (4) - *M* src/bin/latgen-faster-mapped-parallel.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-1> (8) - *M* src/bin/latgen-faster-mapped.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-2> (4) - *M* src/fstbin/fstisstochastic.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-3> (2) - *M* src/fstext/fstext-utils-inl.h <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-4> (25) - *M* src/fstext/fstext-utils.h <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-5> (2) - *M* src/fstext/kaldi-fst-io.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-6> (65) - *M* src/fstext/kaldi-fst-io.h <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-7> (21) - *M* src/gmmbin/gmm-latgen-biglm-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-8> (11) - *M* src/gmmbin/gmm-latgen-faster-parallel.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-9> (5) - *M* src/gmmbin/gmm-latgen-faster-regtree-fmllr.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-10> (4) - *M* src/gmmbin/gmm-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-11> (4) - *M* src/gmmbin/gmm-latgen-map.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-12> (4) - *M* src/gmmbin/gmm-latgen-simple.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-13> (4) - *M* src/nnet2bin/nnet-latgen-faster-parallel.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-14> (6) - *M* src/nnet2bin/nnet-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-15> (4) - *M* src/nnet3bin/nnet3-latgen-faster-looped.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-16> (4) - *M* src/nnet3bin/nnet3-latgen-faster-parallel.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-17> (4) - *M* src/nnet3bin/nnet3-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-18> (4) - *M* src/online2bin/online2-wav-gmm-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-19> (2) - *M* src/online2bin/online2-wav-nnet2-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-20> (2) - *M* src/online2bin/online2-wav-nnet2-latgen-threaded.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-21> (2) - *M* src/online2bin/online2-wav-nnet3-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-22> (2) - *M* src/sgmm2bin/sgmm2-latgen-faster-parallel.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-23> (7) - *M* src/sgmm2bin/sgmm2-latgen-faster.cc <https://github.com/kaldi-asr/kaldi/pull/1661/files#diff-24> (6) Patch Links: - https://github.com/kaldi-asr/kaldi/pull/1661.patch - https://github.com/kaldi-asr/kaldi/pull/1661.diff — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1661>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADJVu7A9crcOGfWPx9_pXWeQq1Y_X1Mzks5r_gPrgaJpZM4NsWeK> . |
Contributor Author
| Hi Dan, |
danpovey reviewed Jun 1, 2017
src/fstext/kaldi-fst-io.h Outdated
| // type VectorFst<StdArc>. If this succeeds, it returns the same pointer; | ||
| // if it fails, it converts the FST type (by creating a new VectorFst<stdArc> | ||
| // initialized by 'fst'), prints a warning, and deletes 'fst'. | ||
| VectorFst<StdArc> *CastToVectorFst(Fst<StdArc> *fst); |
Contributor
There was a problem hiding this comment.
Actually, let's rename this function to CastOrConvertToVectorFst(), which is more informative.
Contributor
There was a problem hiding this comment.
No need to re-est after that change, just make sure it compiles.
Contributor Author
| I have already renamed the function and it can be compiled. |
Contributor
| Thanks, I'll probably merge once the automated checks complete. …On Thu, Jun 1, 2017 at 7:06 PM, LvHang ***@***.***> wrote: I have already renamed the function and it can be compiled. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1661 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADJVu-lJb_ict-RlPzFGuwFe4mlzhiGeks5r_0PtgaJpZM4NsWeK> . |
Contributor
| Thanks! merged. |
kronos-cm added a commit to kronos-cm/kaldi that referenced this pull request Jun 15, 2017
* 'master' of https://github.com/kaldi-asr/kaldi: (140 commits) [egs] Fix failure in multilingual BABEL recipe (regenerate cmvn.scp) (kaldi-asr#1686) [src,scripts,egs] Backstitch code+scripts, and one experiment, will add more later. (kaldi-asr#1605) [egs] CNN+TDNN+LSTM experiments on AMI (kaldi-asr#1685) [egs,scripts,src] Tune image recognition examples; minor small changes. (kaldi-asr#1682) [src] Fix bug in looped computation (kaldi-asr#1673) [build] when installing sequitur and mmseg, look for lib64 as well (thanks: @akshayc11) (kaldi-asr#1677) [src] fix to gst-plugin/Makefile (remove -lkaldi-thread) (kaldi-asr#1680) [src] Cosmetic fixes to usage messages [egs] Fix to some --proportional-shrink related example scripts (kaldi-asr#1674) [build] Fix small bug in configure [scripts] Fix small bug in utils/gen_topo.pl. [scripts] Add python script to convert nnet2 to nnet3 models (kaldi-asr#1611) [doc] Fix typo (kaldi-asr#1669) [src] nnet3: fix small bug in checking code. Thanks: @Maddin2000. [src] Add #include missing from previous commit [src] Fix bug in online2-nnet3 decoding RE dropout+batch-norm (thanks: Wonkyum Lee) [scripts] make errors getting report non-fatal (thx: Miguel Jette); add comment RE dropout proportion [src,scripts] Use ConstFst or decoding (half the memory; slightly faster). (kaldi-asr#1661) [src] keyword search tools: fix Minimize() call, necessary due to OpenFst upgrade (kaldi-asr#1663) [scripts] do not fail if the ivector extractor belongs to different user (kaldi-asr#1662) ...
Skaiste pushed a commit to Skaiste/idlak that referenced this pull request Sep 26, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit. This suggestion is invalid because no changes were made to the code. Suggestions cannot be applied while the pull request is closed. Suggestions cannot be applied while viewing a subset of changes. Only one suggestion per line can be applied in a batch. Add this suggestion to a batch that can be applied as a single commit. Applying suggestions on deleted lines is not supported. You must change the existing code in this line in order to create a valid suggestion. Outdated suggestions cannot be applied. This suggestion has been applied or marked resolved. Suggestions cannot be applied from pending reviews. Suggestions cannot be applied on multi-line comments. Suggestions cannot be applied while the pull request is queued to merge. Suggestion cannot be applied right now. Please check back later.
This PR goes on #1548
This version has removed the FstHolder. That means only "if" branch in *-latgen-faster is fixed in order to speeding up and saving memory about decoding.
@danpovey
The test about nnet3 tdnn training has passed with swbd corpus.
Now, I'm testing the whole pipeline(run.sh) with mini_librispeech corpus. I will notice you when it accomplish.
Bests,
Hang