In the alignment output, there are 2 NSN_S between EVERY sentence. In the transcript...
I guess the main reason is that the model for silence is poor, so it cannot add SIL...
I thought both SIL and NSN are generated by Kaldi, no? I am using the TedLium scripts....
Suppose I have a sentence: Well let me see yes that's it. The actual spoken speech...
Forget about my question. Indeed it is normal for some utterances (e.g. singing)...
Thank you very much indeed. Apart from increasing the value of beam and re-try beam,...
Thank you very much indeed. Apart from increasing the value of beam and re-try beam,...
Thank you very much indeed. Apart from increasing the value of beam and re-try beam,...
Thank you very much indeed. Apart from increasing the value of beam and re-try beam,...
Thank you very much. So beam is like a threshold or something? Will it affect the...
I fed a transcript of some 1000 sentences to Kaldi to do alignment. After finished,...
Regarding <src-dir> the code progresses from tri1 to tri2 to tri3 to dbn_dnn to dbn_dnn_smbr....
Thank you very much. I looked into the nnet/align.sh. It takes in <data-dir>, <lang-dir>,...
I have finished training a DNN model. Now I have a new data with audio and transcript,...
steps/nnet/pretrain_dbn.sh does not have --use-gpu=no option and the inside scripts...
OK I guess one possible source of error comes from: $ cat cmvn_glob_fwd.log nnet-forward...
Would you tell me the meaning of this error? Preparing train/cv lists 31764 exp/dnn4_pretrain-dbn/train.scp...
This Guoguo Chen changes the code to: cat $tmpdir/lexiconp.txt | awk '{print $1}'...
Damn... The 2014 version in Server vs the 2013 version in my Mac
The scripts in utils/prepare_lang.sh cannot write things after the line #0. But this...
The scripts in utils/prepare_lang.sh cannot write things after the line #0. But this...
More precisely, in the server, the words.txt ends with #0 3229 <s> 3230 </s> 3231...
More precisely, in the server, the words.txt ends with #0 3229 <s> 3230 </s> 3231...
More precisely, in the server, the words.txt ends with #0 3229 <s> 3230 </s> 3231...
More precisely, in the server, the words.txt ends with #0 3229 3230 3231 In Mac the...
More precisely, in the server, the words.txt ends with #0 3229 3230 3231 In Mac the...
More precisely, in the server, the words.txt ends with 0 3229 3230 3231 In Mac the...
The words.txt (data/lang and data/lang_test) is different, while all other files...
My scripts run smoothly on a MacBook. But when I run the same script on Amazon server,...
My scripts run smoothly on a MacBook. But when I run the same script on Amazon server,...
I am preparing for a transcript in STM format, but there are some uncertainties on...
I am preparing for a transcript in STM format, but there are some uncertainties on...
I do not want to use lattice but want to directly use the ali.1.gz to get word level...
I do not want to use lattice but want to directly use the ali.1.gz to get word level...
Email sent. Thank you very much.
$ fstinfo G.fst fst type vector arc type standard input symbol table none output...
$ fstinfo G.fst fst type vector arc type standard input symbol table none output...
$ fstinfo G.fst fst type vector arc type standard input symbol table none output...
-rw-r--r-- 1 david staff 66 Dec 2 11:41 G.fst -rw-r--r-- 1 david staff 372290 Dec...
I deleted the whole data directory and ran the program again. ... steps/train_deltas.sh...
My DIC file has < s > SIL and < /s > SIL. Would it cause problem? And I use CMU SLM...
My DIC file has SIL and SIL. Would it cause problem? And I use CMU SLM to generate...
This one? WARNING (fstcomposecontext:ContextFstImpl():fstext/context-fst-inl.h:107)...
I modified the TedLium example to run my own data but came to the following error...
I modified the TedLium example to run my own data but came to the following error...
Thank you very much Daniel. Where can I find this VAD in Kaldi? I saw there are align_si,...
Thank you very much Daniel. I saw there are align_si, align_sgmm2, align_fmllr, etc....
I would like to train a speech detector or VAD in a noisy environment. At first I...
I would like to train a speech detector or VAD in a noisy environment. At first I...
Context is just the consecutive rows of examples in the data. Anyway can we set Context=1...
Dear all, I would like to use the Kaldi DNN implementation but I already have the...
Dear Daniel, The patch works. I have installed Kaldi. Thank you very much. best regards,...
I am trying to install Kaldi on Mac OSX 10.9.4 I got this error when: $ cd kaldi-trunk/tools/...
I am trying to install Kaldi on Mac OSX 10.9.6 I got this error when: $ cd kaldi-trunk/tools/...
I got this error when I make: rfilter1.c:29:5: error: second parameter of 'main'...
I got this error when I make: rfilter1.c:29:5: error: second parameter of 'main'...