Menu
▾
▴
kaldi-developers — Kaldi Developers
You can subscribe to this list here.
| 2011 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(4) |
Jul
|
Aug
|
Sep
(1) |
Oct
(4) |
Nov
(1) |
Dec
(14) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2012 |
Jan
(1) |
Feb
(8) |
Mar
|
Apr
(1) |
May
(3) |
Jun
(13) |
Jul
(7) |
Aug
(11) |
Sep
(6) |
Oct
(14) |
Nov
(16) |
Dec
(1) |
| 2013 |
Jan
(3) |
Feb
(8) |
Mar
(17) |
Apr
(21) |
May
(27) |
Jun
(11) |
Jul
(11) |
Aug
(21) |
Sep
(39) |
Oct
(17) |
Nov
(39) |
Dec
(28) |
| 2014 |
Jan
(36) |
Feb
(30) |
Mar
(35) |
Apr
(17) |
May
(22) |
Jun
(28) |
Jul
(23) |
Aug
(41) |
Sep
(17) |
Oct
(10) |
Nov
(22) |
Dec
(56) |
| 2015 |
Jan
(30) |
Feb
(32) |
Mar
(37) |
Apr
(28) |
May
(79) |
Jun
(18) |
Jul
(35) |
Aug
|
Sep
(1) |
Oct
|
Nov
|
Dec
|
|
From: Daniel P. <dp...@gm...> - 2013-10-29 16:06:35
|
The number of leaves should decrease as you decrease the data, but less than proportionally (e.g. 1/10 the data -> 1/4 the number of leaves). The number of Gaussians per leaf, which is the ratio of tot-num-gauss to num-leaves, should decrease as you decrease the data, e.g. 1/10 the data -> maybe 1/2 the average number of Gaussians per leaf. The total number of Gaussians should decrease in a slightly-less-than-proportional way as you decrease the amount of data. Dan On Tue, Oct 29, 2013 at 12:04 PM, Arif Khan <ife...@gm...> wrote: > Thanks Daniel. > > Do we have any fixed rule, or it is just trying different combinations and > see the results. > Are the No. of leaves and No. of Gaussian dependent on each other OR they > are independent means one can change either of them in any proportions. > > > > Thank you very much. > > Best regards, > Arif > > > On Tue, Oct 29, 2013 at 4:39 PM, Daniel Povey <dp...@gm...> wrote: >> >> If you have only 1 hour of data, you will have to modify the arguments >> to the triphone-training script, i.e. the number of tree leaves and >> number of Gaussians should be reduced. >> Dan >> >> >> On Tue, Oct 29, 2013 at 11:37 AM, Arif Khan <ife...@gm...> wrote: >> > Thanks Daniel for you quick answer. >> > >> > I have another question. I train a model using the WSJ S1 and S5 recipe, >> > and >> > got good results on monophone system than triphone. I have about 1000 >> > utterances, with length of ~1 hr, and about 300 vocabulary size. >> > >> > In theory, the triphone system should perform better than monophone. So >> > I >> > don't know if something is wrong in the tree construction or any thing >> > else >> > could be fixed. >> > >> > Best regards, >> > Arif >> > >> > >> > On Tue, Oct 29, 2013 at 4:17 PM, Daniel Povey <dp...@gm...> wrote: >> >> >> >> Sorry, there is no code per se to get those kind of stats, but you >> >> could perhaps convert the alignments into phone sequences and get the >> >> stats by hand (see ali-to-phones). >> >> >> >> Regarding the tree construction process, there is probably some >> >> documentation on kaldi.sf.net; if there is a particular aspect of that >> >> that is unclear, please let us know; but otherwise, I doubt anyone >> >> has time to respond to your question right now. >> >> >> >> Dan >> >> >> >> >> >> On Tue, Oct 29, 2013 at 10:42 AM, Arif Khan <ife...@gm...> >> >> wrote: >> >> > Dear Kaldi authors, >> >> > >> >> > I want to do some analysis of the training data. Basically I want to >> >> > find >> >> > out the number of phones (mono phone and tri phone) that appeared in >> >> > the >> >> > training set. (relative frequency). Is there any module/script >> >> > available >> >> > for >> >> > it. >> >> > >> >> > Also, I want to find out the tree construction process. I know the >> >> > basics >> >> > from wsj/s5 recipe. But, some more details will be helpful. >> >> > >> >> > >> >> > Best regards, >> >> > Arif >> >> > >> >> > >> >> > >> >> > ------------------------------------------------------------------------------ >> >> > Android is increasing in popularity, but the open development >> >> > platform >> >> > that >> >> > developers love is also attractive to malware creators. Download this >> >> > white >> >> > paper to learn more about secure code signing practices that can help >> >> > keep >> >> > Android apps secure. >> >> > >> >> > >> >> > http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk >> >> > _______________________________________________ >> >> > Kaldi-developers mailing list >> >> > Kal...@li... >> >> > https://lists.sourceforge.net/lists/listinfo/kaldi-developers >> >> > >> > >> > > > |
|
From: Arif K. <ife...@gm...> - 2013-10-29 16:04:38
|
Thanks Daniel. 1. Do we have any fixed rule, or it is just trying different combinations and see the results. 2. Are the No. of leaves and No. of Gaussian dependent on each other OR they are independent means one can change either of them in any proportions. Thank you very much. Best regards, Arif On Tue, Oct 29, 2013 at 4:39 PM, Daniel Povey <dp...@gm...> wrote: > If you have only 1 hour of data, you will have to modify the arguments > to the triphone-training script, i.e. the number of tree leaves and > number of Gaussians should be reduced. > Dan > > > On Tue, Oct 29, 2013 at 11:37 AM, Arif Khan <ife...@gm...> wrote: > > Thanks Daniel for you quick answer. > > > > I have another question. I train a model using the WSJ S1 and S5 recipe, > and > > got good results on monophone system than triphone. I have about 1000 > > utterances, with length of ~1 hr, and about 300 vocabulary size. > > > > In theory, the triphone system should perform better than monophone. So > I > > don't know if something is wrong in the tree construction or any thing > else > > could be fixed. > > > > Best regards, > > Arif > > > > > > On Tue, Oct 29, 2013 at 4:17 PM, Daniel Povey <dp...@gm...> wrote: > >> > >> Sorry, there is no code per se to get those kind of stats, but you > >> could perhaps convert the alignments into phone sequences and get the > >> stats by hand (see ali-to-phones). > >> > >> Regarding the tree construction process, there is probably some > >> documentation on kaldi.sf.net; if there is a particular aspect of that > >> that is unclear, please let us know; but otherwise, I doubt anyone > >> has time to respond to your question right now. > >> > >> Dan > >> > >> > >> On Tue, Oct 29, 2013 at 10:42 AM, Arif Khan <ife...@gm...> > wrote: > >> > Dear Kaldi authors, > >> > > >> > I want to do some analysis of the training data. Basically I want to > >> > find > >> > out the number of phones (mono phone and tri phone) that appeared in > the > >> > training set. (relative frequency). Is there any module/script > available > >> > for > >> > it. > >> > > >> > Also, I want to find out the tree construction process. I know the > >> > basics > >> > from wsj/s5 recipe. But, some more details will be helpful. > >> > > >> > > >> > Best regards, > >> > Arif > >> > > >> > > >> > > ------------------------------------------------------------------------------ > >> > Android is increasing in popularity, but the open development platform > >> > that > >> > developers love is also attractive to malware creators. Download this > >> > white > >> > paper to learn more about secure code signing practices that can help > >> > keep > >> > Android apps secure. > >> > > >> > > http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk > >> > _______________________________________________ > >> > Kaldi-developers mailing list > >> > Kal...@li... > >> > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > >> > > > > > > |
|
From: Daniel P. <dp...@gm...> - 2013-10-29 15:39:35
|
If you have only 1 hour of data, you will have to modify the arguments to the triphone-training script, i.e. the number of tree leaves and number of Gaussians should be reduced. Dan On Tue, Oct 29, 2013 at 11:37 AM, Arif Khan <ife...@gm...> wrote: > Thanks Daniel for you quick answer. > > I have another question. I train a model using the WSJ S1 and S5 recipe, and > got good results on monophone system than triphone. I have about 1000 > utterances, with length of ~1 hr, and about 300 vocabulary size. > > In theory, the triphone system should perform better than monophone. So I > don't know if something is wrong in the tree construction or any thing else > could be fixed. > > Best regards, > Arif > > > On Tue, Oct 29, 2013 at 4:17 PM, Daniel Povey <dp...@gm...> wrote: >> >> Sorry, there is no code per se to get those kind of stats, but you >> could perhaps convert the alignments into phone sequences and get the >> stats by hand (see ali-to-phones). >> >> Regarding the tree construction process, there is probably some >> documentation on kaldi.sf.net; if there is a particular aspect of that >> that is unclear, please let us know; but otherwise, I doubt anyone >> has time to respond to your question right now. >> >> Dan >> >> >> On Tue, Oct 29, 2013 at 10:42 AM, Arif Khan <ife...@gm...> wrote: >> > Dear Kaldi authors, >> > >> > I want to do some analysis of the training data. Basically I want to >> > find >> > out the number of phones (mono phone and tri phone) that appeared in the >> > training set. (relative frequency). Is there any module/script available >> > for >> > it. >> > >> > Also, I want to find out the tree construction process. I know the >> > basics >> > from wsj/s5 recipe. But, some more details will be helpful. >> > >> > >> > Best regards, >> > Arif >> > >> > >> > ------------------------------------------------------------------------------ >> > Android is increasing in popularity, but the open development platform >> > that >> > developers love is also attractive to malware creators. Download this >> > white >> > paper to learn more about secure code signing practices that can help >> > keep >> > Android apps secure. >> > >> > http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk >> > _______________________________________________ >> > Kaldi-developers mailing list >> > Kal...@li... >> > https://lists.sourceforge.net/lists/listinfo/kaldi-developers >> > > > |
|
From: Arif K. <ife...@gm...> - 2013-10-29 15:37:40
|
Thanks Daniel for you quick answer. I have another question. I train a model using the WSJ S1 and S5 recipe, and got good results on monophone system than triphone. I have about 1000 utterances, with length of ~1 hr, and about 300 vocabulary size. In theory, the triphone system should perform better than monophone. So I don't know if something is wrong in the tree construction or any thing else could be fixed. Best regards, Arif On Tue, Oct 29, 2013 at 4:17 PM, Daniel Povey <dp...@gm...> wrote: > Sorry, there is no code per se to get those kind of stats, but you > could perhaps convert the alignments into phone sequences and get the > stats by hand (see ali-to-phones). > > Regarding the tree construction process, there is probably some > documentation on kaldi.sf.net; if there is a particular aspect of that > that is unclear, please let us know; but otherwise, I doubt anyone > has time to respond to your question right now. > > Dan > > > On Tue, Oct 29, 2013 at 10:42 AM, Arif Khan <ife...@gm...> wrote: > > Dear Kaldi authors, > > > > I want to do some analysis of the training data. Basically I want to find > > out the number of phones (mono phone and tri phone) that appeared in the > > training set. (relative frequency). Is there any module/script available > for > > it. > > > > Also, I want to find out the tree construction process. I know the basics > > from wsj/s5 recipe. But, some more details will be helpful. > > > > > > Best regards, > > Arif > > > > > ------------------------------------------------------------------------------ > > Android is increasing in popularity, but the open development platform > that > > developers love is also attractive to malware creators. Download this > white > > paper to learn more about secure code signing practices that can help > keep > > Android apps secure. > > > http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk > > _______________________________________________ > > Kaldi-developers mailing list > > Kal...@li... > > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > > > |
|
From: Daniel P. <dp...@gm...> - 2013-10-29 15:18:02
|
Sorry, there is no code per se to get those kind of stats, but you could perhaps convert the alignments into phone sequences and get the stats by hand (see ali-to-phones). Regarding the tree construction process, there is probably some documentation on kaldi.sf.net; if there is a particular aspect of that that is unclear, please let us know; but otherwise, I doubt anyone has time to respond to your question right now. Dan On Tue, Oct 29, 2013 at 10:42 AM, Arif Khan <ife...@gm...> wrote: > Dear Kaldi authors, > > I want to do some analysis of the training data. Basically I want to find > out the number of phones (mono phone and tri phone) that appeared in the > training set. (relative frequency). Is there any module/script available for > it. > > Also, I want to find out the tree construction process. I know the basics > from wsj/s5 recipe. But, some more details will be helpful. > > > Best regards, > Arif > > ------------------------------------------------------------------------------ > Android is increasing in popularity, but the open development platform that > developers love is also attractive to malware creators. Download this white > paper to learn more about secure code signing practices that can help keep > Android apps secure. > http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > |
|
From: Arif K. <ife...@gm...> - 2013-10-29 14:42:46
|
Dear Kaldi authors, I want to do some analysis of the training data. Basically I want to find out the number of phones (mono phone and tri phone) that appeared in the training set. (relative frequency). Is there any module/script available for it. Also, I want to find out the tree construction process. I know the basics from wsj/s5 recipe. But, some more details will be helpful. Best regards, Arif |
|
From: Daniel P. <dp...@gm...> - 2013-10-20 16:02:28
|
It looks like you are using older or "stable" code. Try the "trunk" code, which does failover to Sourceforge if the primary locations of those software packages are not accessible from your location. Dan On Sat, Oct 19, 2013 at 7:10 AM, bengu wu <wbg...@gm...> wrote: > dear kaldi-developers, > first,let me introduce myself.i am a master student of China.now i > learn the speech recognition. > when i install kaldi from the website,i have the failure.the failure is > next: > Install summary: > sph2pipe:Success > ATLAS: Success [note: we install just the headers; do ./install_atlas.sh > if ../src/configure fails.] > CLAPACK: Success > irstlm: Failure [optional anyway] > sclite: Failure [optional anyway.. see INSTALL for more help] > openfst: Success > > i try some times,but it also fail .can you help me?thank you very much. > best wishes > yours > ben > > > ------------------------------------------------------------------------------ > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most > from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > |
|
From: bengu wu <wbg...@gm...> - 2013-10-19 11:10:35
|
dear kaldi-developers,
first,let me introduce myself.i am a master student of China.now i
learn the speech recognition.
when i install kaldi from the website,i have the failure.the failure is
next:
Install summary:
sph2pipe:Success
ATLAS: Success [note: we install just the headers; do ./install_atlas.sh
if ../src/configure fails.]
CLAPACK: Success
irstlm: Failure [optional anyway]
sclite: Failure [optional anyway.. see INSTALL for more help]
openfst: Success
i try some times,but it also fail .can you help me?thank you very much.
best wishes
yours
ben
|
|
From: Arnab G. <ar...@gm...> - 2013-10-15 08:00:12
|
Hi all, there are two open positions at the University of Edinburgh. For the ASR position, experience with Kaldi will be highly valued. Please forward this to people who may be interested. -Arnab Postdoctoral Research Positions The Centre for Speech Technology Research University of Edinburgh http://www.cstr.ed.ac.uk/opportunities/ Closing date: 18th November 2013 We have open postdoctoral research positions in speech recognition and speech synthesis, which are part of the large UK project "Natural Speech Technology", http://www.natural-speech-technology.org. The topics of research to be undertaken in these positions is flexible. We encourage applicants with their own research agenda, provided that it fits within the objective of the project, which is to advance the state of the art in speech technology by making it more natural, approaching human levels of reliability, adaptability, and conversational richness. SPEECH SYNTHESIS: Our specific interests in speech synthesis include - but are not limited to - the following: * Machine learning for vocoding. * Fluent speech synthesis. * Shallow stochastic natural language generation to improve fluency. * Beyond decision tree parameter tying, including neural network approaches or tree intersect models. * Use of synthetic speech in assistive technologies. SPEECH RECOGNITION: Our specific interests in speech recognition include - but are not limited to - the following: * Wide domain coverage and models which make use of rich contexts. * Cross-lingual speech recognition. * Neural network models. * Adaptation and canonical modelling techniques for acoustic or language modelling. * Distant speech recognition. * Approaches based on models incorporating articulatory data. The Centre for Speech Technology Research (CSTR) is an exciting, vibrant, interdisciplinary research centre and a great place to work. We are part of the University of Edinburgh (QS world ranking 17th) linking the world-class subject areas of informatics / computer science (QS world ranking 15th) and linguistics (QS world ranking 5th) Founded in 1984, CSTR is concerned with research in all areas of speech technology including speech recognition, speech synthesis, speech signal processing, information access, multimodal interfaces and dialogue systems. We have many significant collaborations with the wider community of researchers in speech science, language, cognition and machine learning for which Edinburgh is renowned, and a wide network of collaborators across the globe. For further details, and links to the online application procedure please visithttp://www.cstr.ed.ac.uk/opportunities/ Informal enquiries about these positions should be made to Prof Steve Renals (s.r...@ed...) or to Prof Simon King (Sim...@ed...). |
|
From: Ben J. <be...@ne...> - 2013-10-12 00:56:27
|
Just a quick followup on this.... the parameter changes worked. The training seems to be stable now. Ben On Mon, Sep 2, 2013 at 8:12 PM, Ben Jiang <be...@ne...> wrote: > Got it. Really appreciate the help here! I'll report any findings back > here. > > Ben > > > On Mon, Sep 2, 2013 at 8:06 PM, Daniel Povey <dp...@gm...> wrote: > >> Sorry, in rm/s5, it's local/run_nnet2.sh, in wsj/s5 it's >> local/run_nnet_cpu.sh >> Dan >> >> >> On Mon, Sep 2, 2013 at 8:04 PM, Ben Jiang <be...@ne...> wrote: >> > Ok, got it. Let me try 200k first. >> > >> > I just updated the trunk, but couldn't find run_nnet2.sh. Is it >> supposed to >> > be in wsj/s5/local/? >> > >> > >> > Thanks >> > Ben >> > >> > >> > >> > On Mon, Sep 2, 2013 at 7:52 PM, Daniel Povey <dp...@gm...> wrote: >> >> >> >> That log-prob per frame if -7.31 is too low, it should be something >> >> like -2, no lower-- maybe -3 on the 1st iteration. The size of your >> >> training data does not matter, what matters is the #samples you >> >> process per iteration. Maybe try reducing it from 400k (the default, >> >> I think) to 200k. Or use the newer example scripts where I think that >> >> is the default. (if you update the trunk and look at the example >> >> scripts run_nnet2.sh, you'll see what I mean). >> >> >> >> But definitely something is wrong here. >> >> >> >> Dan >> >> >> >> >> >> On Mon, Sep 2, 2013 at 7:47 PM, Ben Jiang <be...@ne...> wrote: >> >> > The nonlinearaty type should be the default in train_nnet_cpu.sh, >> which >> >> > should be tanh. The log-prob doesn't look too bad. Below is the >> output >> >> > from >> >> > a run that actually succeeded: >> >> > LOG >> >> > >> (nnet-train-parallel:DoBackpropParallel():nnet-update-parallel.cc:179) >> >> > Did backprop on 399889 examples, average log-prob per frame is >> -7.31817 >> >> > >> >> > The learning rates are 0.01 initial and 0.001 final. I kind of used >> the >> >> > value from swbd, but maybe my training data is quite bigger than >> swbd. >> >> > I >> >> > previously tried 0.001 and 0.0001, which also failed due to an error >> of >> >> > "Cannot invert: matrix is singular", but I didn't have debug on back >> >> > then, >> >> > so it's probably the same issue. Maybe I should try even smaller, >> such >> >> > as >> >> > 0.0001 and 0.00001? >> >> > >> >> > >> >> > Ben >> >> > >> >> > >> >> > >> >> > On Mon, Sep 2, 2013 at 6:55 PM, Daniel Povey <dp...@gm...> >> wrote: >> >> >> >> >> >> I think the underlying cause is instability in the training, causing >> >> >> the derivatives to become too large. This is something that >> commonly >> >> >> happens in neural net training, and the solution is generally to >> >> >> decrease the learning rate. What nonlinearity type are you using? >> >> >> And do the log-probs printed out in train.*.log or >> compute_prob_*.log >> >> >> get very negative? >> >> >> >> >> >> Unbounded nonlinearities such as ReLUs are more susceptible to this >> >> >> instability. >> >> >> Dan >> >> >> >> >> >> >> >> >> On Mon, Sep 2, 2013 at 6:50 PM, Ben Jiang <be...@ne...> wrote: >> >> >> > I see. Thanks for the fast response, Dan. >> >> >> > >> >> >> > So any idea on this "random" error I am stuck with at pass 27? I >> >> >> > have >> >> >> > pasted the stacktrace below. This error doesn't always happen, >> even >> >> >> > after >> >> >> > I removed the randomness introduced in the input mdl and shuffled >> >> >> > egs. >> >> >> > (eg, >> >> >> > save the input mdl and shuffled egs to files and re-run the failed >> >> >> > nnet-train-parallel from those files in debugger). The re-run >> would >> >> >> > sometimes fail and sometimes succeed. >> >> >> > >> >> >> > Anyway, I was able catch the error in my debugger and examine the >> >> >> > variables. >> >> >> > I think the reason is that the deriv variable in >> >> >> > NnetUpdater::Backprop() >> >> >> > contains some "bad" value, such as 1.50931703e+20. This caused >> the >> >> >> > trace of >> >> >> > the matrix to become infinite and in turn cause the p_trace to >> become >> >> >> > 0 >> >> >> > and >> >> >> > fail the assert. I probably need more time to see how this value >> got >> >> >> > in >> >> >> > there, but again, since the exact re-run would pass sometimes, >> it's >> >> >> > kind >> >> >> > of >> >> >> > hard to debug. >> >> >> > >> >> >> > Any idea? >> >> >> > >> >> >> > Here's the stacktrace: >> >> >> > =============================== >> >> >> > KALDI_ASSERT: at >> >> >> > >> >> >> > >> >> >> > >> nnet-train-parallel:PreconditionDirectionsAlphaRescaled:nnet-precondition.cc:128, >> >> >> > failed: p_trace != 0.0 >> >> >> > Stack trace is: >> >> >> > kaldi::KaldiGetStackTrace() >> >> >> > kaldi::KaldiAssertFailure_(char const*, char const*, int, char >> >> >> > const*) >> >> >> > >> >> >> > >> >> >> > >> kaldi::nnet2::PreconditionDirectionsAlphaRescaled(kaldi::MatrixBase<float> >> >> >> > const&, double, kaldi::MatrixBase<float>*) >> >> >> > >> >> >> > >> >> >> > >> kaldi::nnet2::AffineComponentPreconditioned::Update(kaldi::MatrixBase<float> >> >> >> > const&, kaldi::MatrixBase<float> const&) >> >> >> > kaldi::nnet2::AffineComponent::Backprop(kaldi::MatrixBase<float> >> >> >> > const&, >> >> >> > kaldi::MatrixBase<float> const&, kaldi::MatrixBase<float> const&, >> >> >> > int, >> >> >> > kaldi::nnet2::Component*, kaldi::Matrix<float>*) const >> >> >> > >> >> >> > >> >> >> > >> kaldi::nnet2::NnetUpdater::Backprop(std::vector<kaldi::nnet2::NnetTrainingExample, >> >> >> > std::allocator<kaldi::nnet2::NnetTrainingExample> > const&, >> >> >> > kaldi::Matrix<float>*) >> >> >> > >> >> >> > >> >> >> > >> kaldi::nnet2::NnetUpdater::ComputeForMinibatch(std::vector<kaldi::nnet2::NnetTrainingExample, >> >> >> > std::allocator<kaldi::nnet2::NnetTrainingExample> > const&) >> >> >> > kaldi::nnet2::DoBackprop(kaldi::nnet2::Nnet const&, >> >> >> > std::vector<kaldi::nnet2::NnetTrainingExample, >> >> >> > std::allocator<kaldi::nnet2::NnetTrainingExample> > const&, >> >> >> > kaldi::nnet2::Nnet*) >> >> >> > kaldi::nnet2::DoBackpropParallelClass::operator()() >> >> >> > kaldi::MultiThreadable::run(void*) >> >> >> > >> >> >> > Ben >> >> >> > >> >> >> > >> >> >> > On Mon, Sep 2, 2013 at 6:25 PM, Daniel Povey <dp...@gm...> >> >> >> > wrote: >> >> >> >> >> >> >> >> That's how it's supposed to be-- AFAIK that's basically the >> point of >> >> >> >> Hogwild, that you allow these kinds of updates and accept the >> >> >> >> possibility that due to race conditions you will occasionally >> lose a >> >> >> >> bit of date. The parameters only change slightly on the >> timescales >> >> >> >> that these different threads access them. >> >> >> >> Dan >> >> >> >> >> >> >> >> >> >> >> >> On Mon, Sep 2, 2013 at 6:01 PM, Ben Jiang <be...@ne...> >> wrote: >> >> >> >> > Hi all, >> >> >> >> > >> >> >> >> > While hunting some random error from nnet-train-parallel, I >> >> >> >> > noticed >> >> >> >> > the >> >> >> >> > nnet_to_update is shared among the threads, but there is no >> >> >> >> > synchronization >> >> >> >> > checks when updating the components in the threads. I still >> >> >> >> > haven't >> >> >> >> > gone >> >> >> >> > too deep in the code yet, but should there be synchronization >> >> >> >> > checks? >> >> >> >> > >> >> >> >> > For example, the deriv variable in NnetUpdater::Backprop() is >> >> >> >> > updated >> >> >> >> > and >> >> >> >> > passed between the components. Could this be an issue if the >> >> >> >> > components >> >> >> >> > are >> >> >> >> > being updated by other threads? >> >> >> >> > >> >> >> >> > >> >> >> >> > Or am I missing something totally? >> >> >> >> > >> >> >> >> > >> >> >> >> > -- >> >> >> >> > Thanks >> >> >> >> > Ben >> >> >> >> > >> >> >> >> > >> >> >> >> > >> >> >> >> > >> >> >> >> > >> ------------------------------------------------------------------------------ >> >> >> >> > Learn the latest--Visual Studio 2012, SharePoint 2013, SQL >> 2012, >> >> >> >> > more! >> >> >> >> > Discover the easy way to master current and previous Microsoft >> >> >> >> > technologies >> >> >> >> > and advance your career. Get an incredible 1,500+ hours of >> >> >> >> > step-by-step >> >> >> >> > tutorial videos with LearnDevNow. Subscribe today and save! >> >> >> >> > >> >> >> >> > >> >> >> >> > >> >> >> >> > >> http://pubads.g.doubleclick.net/gampad/clk?id=58040911&iu=/4140/ostg.clktrk >> >> >> >> > _______________________________________________ >> >> >> >> > Kaldi-developers mailing list >> >> >> >> > Kal...@li... >> >> >> >> > https://lists.sourceforge.net/lists/listinfo/kaldi-developers >> >> >> >> > >> >> >> > >> >> >> > >> >> >> > >> >> >> > >> >> >> > -- >> >> >> > >> >> >> > -- >> >> >> > Thanks >> >> >> > Ben Jiang >> >> >> > >> >> >> > Co-Founder/Principal/CTO >> >> >> > Nexiwave.com >> >> >> > Tel: 226-975-2172 / 617-245-0916 >> >> >> > "Confidential & Privileged: This email message is for the sole >> use of >> >> >> > the >> >> >> > intended recipient(s) and may contain confidential and privileged >> >> >> > information. Any unauthorized review, use, disclosure or >> distribution >> >> >> > is >> >> >> > prohibited. if you are not the intended recipient, please contact >> the >> >> >> > sender >> >> >> > by reply email and destroy all copies of the original message.” >> >> > >> >> > >> >> > >> >> > >> >> > -- >> >> > >> >> > -- >> >> > Thanks >> >> > Ben Jiang >> >> > >> >> > Co-Founder/Principal/CTO >> >> > Nexiwave.com >> >> > Tel: 226-975-2172 / 617-245-0916 >> >> > "Confidential & Privileged: This email message is for the sole use of >> >> > the >> >> > intended recipient(s) and may contain confidential and privileged >> >> > information. Any unauthorized review, use, disclosure or >> distribution is >> >> > prohibited. if you are not the intended recipient, please contact the >> >> > sender >> >> > by reply email and destroy all copies of the original message.” >> > >> > >> > >> > >> > -- >> > >> > -- >> > Thanks >> > Ben Jiang >> > >> > Co-Founder/Principal/CTO >> > Nexiwave.com >> > Tel: 226-975-2172 / 617-245-0916 >> > "Confidential & Privileged: This email message is for the sole use of >> the >> > intended recipient(s) and may contain confidential and privileged >> > information. Any unauthorized review, use, disclosure or distribution is >> > prohibited. if you are not the intended recipient, please contact the >> sender >> > by reply email and destroy all copies of the original message.” >> > > > > -- > > -- > Thanks > Ben Jiang > > Co-Founder/Principal/CTO > Nexiwave.com > Tel: 226-975-2172 / 617-245-0916 > "Confidential & Privileged: This email message is for the sole use of the > intended recipient(s) and may contain confidential and privileged > information. Any unauthorized review, use, disclosure or distribution is > prohibited. if you are not the intended recipient, please contact the > sender by reply email and destroy all copies of the original message.” > |
|
From: Jan T. <jt...@gm...> - 2013-09-30 12:45:09
|
Hi, 1) you don't have to build the ATLAS library, as Dan said, it is sufficient to install whatever package your system distribution provides. Atlas is sometimes quite hard to build because it tries to optimize everything for the machine on which it's being compiled. It also means that when you have some specific machine configuration, the automatic build fails. 2) Kaldi supports OpenBLAS and/or Intel MKL already. You just have to install either one of those and call the configure script with the correct parameters. Configure --help lists these parameters Y. On Sep 30, 2013 1:16 AM, "殷永光" <yg...@if...> wrote: > ** > Thanks for you advice, I am still in teh trouble of building. > But Can you release a trunk version which uses OpenBLAS instead of ATLAS? > > Because I want to use DNN training in Kaldi > and current trunk verion is too hard to build. > Thanks > best regard > Yongguang Yin > ------------------------------ > 平嵌识别 殷永光 > 2013-09-30 > ------------------------------ > *发件人：* Daniel Povey > *发送时间：* 2013-09-27 21:54:21 > *收件人：* 殷永光 > *抄送：* kaldi-developers > *主题：* Re: [Kaldi-developers] What's wrong happens when I install ATLAS > libray? > Building ATLAS is complex. I suggest you just use whatever ATLAS > package comes with your system, i.e. have the sysadmin install the > package, or try OpenBLAS. > Dan > > > On Fri, Sep 27, 2013 at 4:20 AM, 殷永光 <yg...@if... > wrote: > > Bear kal...@li...: > > > when I run the script configure in src , it told me that program didn't find libatlas.so, > > and I shloud run the script install_atlas.sh in directory tools. > > when I run this script install_atlas.sh , error happens. > > the log is > > " > > > ./xconfig -d s /home/ygyin/kaldi/tools/ATLAS/build/../ -d b /home/ygyin/kaldi/tools/ATLAS/build > > ERROR: enum fam=3, chip=2, mach=0 > > make[3]:***[atlas_run] ERROR 44 > > make[2]:***[IRunArchInfo_x86] ERROR 2 > > ERROR: enum fam=3, chip=2, mach=0 > > make[3]:***[atlas_run] ERROR 44 > > make[2]:***[IRunArchInfo_x86] ERROR 2 > > ERROR: enum fam=3, chip=2, mach=0 > > make[3]:***[atlas_run] ERROR 44 > > make[2]:***[IRunArchInfo_x86] ERROR 2 > > ERROR: enum fam=3, chip=2, mach=0 > > make[3]:***[atlas_run] ERROR 44 > > make[2]:***[IRunArchInfo_x86] ERROR 2 > > rm -f config1.out > > ............. > > " > > I really don't known what should I do? Can you help me? > > Thanks > > best wishes! > > Yongguang Yin > > -------------- > > 殷永光 > > 2013-09-27 > > > ------------------------------------------------------------------------------ > > October Webinars: Code for Performance > > Free Intel webinars can help you accelerate application performance. > > > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from > > the latest Intel processors and coprocessors. See abstracts and register > > > > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > > _______________________________________________ > > Kaldi-developers mailing list > > Kal...@li... > > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > > > ------------------------------------------------------------------------------ > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most > from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > > |
|
From: Vassil P. <vas...@gm...> - 2013-09-30 09:48:44
|
Hi, regarding OpenBLAS, as far as I know it's already supported in the current trunk. Did you have a look at http://kaldi.sourceforge.net/matrixwrap.html ? Vassil On Mon, Sep 30, 2013 at 8:16 AM, 殷永光 <yg...@if...> wrote: > Thanks for you advice, I am still in teh trouble of building. > But Can you release a trunk version which uses OpenBLAS instead of ATLAS? > > Because I want to use DNN training in Kaldi > and current trunk verion is too hard to build. > Thanks > best regard > Yongguang Yin > ________________________________ > 平嵌识别 殷永光 > 2013-09-30 > ________________________________ > 发件人： Daniel Povey > 发送时间： 2013-09-27 21:54:21 > 收件人： 殷永光 > 抄送： kaldi-developers > 主题： Re: [Kaldi-developers] What's wrong happens when I install ATLAS libray? > Building ATLAS is complex. I suggest you just use whatever ATLAS > package comes with your system, i.e. have the sysadmin install the > package, or try OpenBLAS. > Dan > > > On Fri, Sep 27, 2013 at 4:20 AM, 殷永光 <yg...@if... > wrote: >> Bear kal...@li...: >> when I run the script configure in src , it told me that program >> didn't find libatlas.so, >> and I shloud run the script install_atlas.sh in directory tools. >> when I run this script install_atlas.sh , error happens. >> the log is >> " >> ./xconfig -d s /home/ygyin/kaldi/tools/ATLAS/build/../ -d b >> /home/ygyin/kaldi/tools/ATLAS/build >> ERROR: enum fam=3, chip=2, mach=0 >> make[3]:***[atlas_run] ERROR 44 >> make[2]:***[IRunArchInfo_x86] ERROR 2 >> ERROR: enum fam=3, chip=2, mach=0 >> make[3]:***[atlas_run] ERROR 44 >> make[2]:***[IRunArchInfo_x86] ERROR 2 >> ERROR: enum fam=3, chip=2, mach=0 >> make[3]:***[atlas_run] ERROR 44 >> make[2]:***[IRunArchInfo_x86] ERROR 2 >> ERROR: enum fam=3, chip=2, mach=0 >> make[3]:***[atlas_run] ERROR 44 >> make[2]:***[IRunArchInfo_x86] ERROR 2 >> rm -f config1.out >> ............. >> " >> I really don't known what should I do? Can you help me? >> Thanks >> best wishes! >> Yongguang Yin >> -------------- >> 殷永光 >> 2013-09-27 >> >> ------------------------------------------------------------------------------ >> October Webinars: Code for Performance >> Free Intel webinars can help you accelerate application performance. >> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most >> from >> the latest Intel processors and coprocessors. See abstracts and register >> > >> >> http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-developers mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-developers > > ------------------------------------------------------------------------------ > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most > from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > |
|
From: Ho Y. C. <ric...@gm...> - 2013-09-30 05:28:57
|
Hi Rafi, FYI. Ubuntu 12.04 and CentOS 6.4 are stable in Kaldi. For Neural Network training with GPU, CUDA 5.0 is stable. 4GB or 8GB memory can handle small size language model (say less than 100MB text size arpa format file) in Kaldi but probably not enough for efficient running with bigger language model (say with size larger than 400MB-500MB text size arpa format file together with 100K wordlist) during some resources preparation (e.g. some finite state transducer operations like arpa2fst together with fst composition etc., which we saw at least 10 times order more memory consumed compare to the arpa format language model file size). -Ricky Message: 1 > Date: Mon, 23 Sep 2013 10:44:47 +0100 > From: Arnab Ghoshal <ar...@gm...> > Subject: Re: [Kaldi-developers] hardware requirements > To: rafi mees <ra...@ec...> > Cc: "kal...@li..." > <kal...@li...> > Message-ID: > <CABPTVix1sQc58mbD=4= > Ei6...@ma...> > Content-Type: text/plain; charset=ISO-8859-1 > > Hi Rafi- > > Kaldi will run with the specs you mention. But the type of datasets > you plan to use will really determine the type and amount of hardware > that will be suitable. For example, if you are training/recognizing > some tens of hours of speech then you may want to run at least 5-10 > parallel processes, or 30-50 processes when training on a few hundred > hours. For training neural networks a GPU is highly recommended (I > normally use GTX 680/690). > > The following is from Chris Kermorvant in response to a different email: > "I have configured an automatic build of Kaldi which runs the > egs/rm/s5 recipe every day. > It runs on a Intel(R) Xeon(R) CPU E5-2630 0 @ 2.30GHz, 12*2 CPU, 64 Gb > memory. > The current time for running this recipe on this computer is 55 minutes." > > Hope this helps. > -Arnab > > > On Tue, Sep 17, 2013 at 11:53 AM, rafi mees <ra...@ec...> wrote: > > Hello. > > > > my name is rafi i am part of IT staff and our development dept. requires > > Kaldi. > > > > if you can advise us on what kind of hardware is to be used: > > > > cpu : dual, quad > > memory: 4, 8, 16 GB > > system: UBUNTU 12.10 or later. > > > > might i add that this would be installed on Virtual machine > > > > Thanks > > > > Rafi Mees > > ECB IT Dept. > > > > > > > > > > > > > ************************************************************************************ > > This footnote confirms that this email message has been scanned by > > PineApp Mail-SeCure for the presence of malicious code, vandals & > computer viruses. > > > ************************************************************************************ > > > > > > > > > > > ------------------------------------------------------------------------------ > > LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99! > > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, > SharePoint > > 2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack > includes > > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > > > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > > _______________________________________________ > > Kaldi-developers mailing list > > Kal...@li... > > https://lists.sourceforge.net/lists/listinfo/kaldi-developers > > > |
|
From: 殷. <yg...@if...> - 2013-09-30 05:16:37
|
Thanks for you advice, I am still in teh trouble of building. But Can you release a trunk version which uses OpenBLAS instead of ATLAS? Because I want to use DNN training in Kaldi and current trunk verion is too hard to build. Thanks best regard Yongguang Yin 平嵌识别 殷永光 2013-09-30 发件人： Daniel Povey 发送时间： 2013-09-27 21:54:21 收件人： 殷永光 抄送： kaldi-developers 主题： Re: [Kaldi-developers] What's wrong happens when I install ATLAS libray? Building ATLAS is complex. I suggest you just use whatever ATLAS package comes with your system, i.e. have the sysadmin install the package, or try OpenBLAS. Dan On Fri, Sep 27, 2013 at 4:20 AM, 殷永光 <yg...@if... > wrote: > Bear kal...@li...: > when I run the script configure in src , it told me that program didn't find libatlas.so, > and I shloud run the script install_atlas.sh in directory tools. > when I run this script install_atlas.sh , error happens. > the log is > " > ./xconfig -d s /home/ygyin/kaldi/tools/ATLAS/build/../ -d b /home/ygyin/kaldi/tools/ATLAS/build > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > rm -f config1.out > ............. > " > I really don't known what should I do? Can you help me? > Thanks > best wishes! > Yongguang Yin > -------------- > 殷永光 > 2013-09-27 > ------------------------------------------------------------------------------ > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
|
From: Shi Z. <shi...@cn...> - 2013-09-30 03:20:49
|
========================================================= Ziqiang Shi（石自强） Information Technology Laboratory Fujitsu Research & Development Center Co., Ltd. 15/F, Tower A, Ocean International Center, No.56 Dong Si Huan Zhong Rd, Chaoyang District, Beijing, 100025, P.R. China TEL: 010-59691594 转 5756 PHONE: +86-13621160486 E-Mail: shi...@cn... <mailto:li...@cn...> ========================================================= |
|
From: Daniel P. <dp...@gm...> - 2013-09-27 13:54:13
|
Building ATLAS is complex. I suggest you just use whatever ATLAS package comes with your system, i.e. have the sysadmin install the package, or try OpenBLAS. Dan On Fri, Sep 27, 2013 at 4:20 AM, 殷永光 <yg...@if...> wrote: > Bear kal...@li...: > when I run the script configure in src , it told me that program didn't find libatlas.so, > and I shloud run the script install_atlas.sh in directory tools. > when I run this script install_atlas.sh , error happens. > the log is > " > ./xconfig -d s /home/ygyin/kaldi/tools/ATLAS/build/../ -d b /home/ygyin/kaldi/tools/ATLAS/build > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > ERROR: enum fam=3, chip=2, mach=0 > make[3]:***[atlas_run] ERROR 44 > make[2]:***[IRunArchInfo_x86] ERROR 2 > rm -f config1.out > ............. > " > I really don't known what should I do? Can you help me? > Thanks > best wishes! > Yongguang Yin > -------------- > 殷永光 > 2013-09-27 > ------------------------------------------------------------------------------ > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
|
From: 殷. <yg...@if...> - 2013-09-27 08:40:25
|
Bear kal...@li...: when I run the script configure in src , it told me that program didn't find libatlas.so, and I shloud run the script install_atlas.sh in directory tools. when I run this script install_atlas.sh , error happens. the log is " ./xconfig -d s /home/ygyin/kaldi/tools/ATLAS/build/../ -d b /home/ygyin/kaldi/tools/ATLAS/build ERROR: enum fam=3, chip=2, mach=0 make[3]:***[atlas_run] ERROR 44 make[2]:***[IRunArchInfo_x86] ERROR 2 ERROR: enum fam=3, chip=2, mach=0 make[3]:***[atlas_run] ERROR 44 make[2]:***[IRunArchInfo_x86] ERROR 2 ERROR: enum fam=3, chip=2, mach=0 make[3]:***[atlas_run] ERROR 44 make[2]:***[IRunArchInfo_x86] ERROR 2 ERROR: enum fam=3, chip=2, mach=0 make[3]:***[atlas_run] ERROR 44 make[2]:***[IRunArchInfo_x86] ERROR 2 rm -f config1.out ............. " I really don't known what should I do? Can you help me? Thanks best wishes! Yongguang Yin -------------- 殷永光 2013-09-27 |
|
From: Daniel P. <dp...@gm...> - 2013-09-25 19:58:42
|
It's out of date-- I thought I had someone fix it, but I guess that didn't happen. Dan On Wed, Sep 25, 2013 at 3:57 PM, Eric Riebling <er...@cs...> wrote: > From the run.sh recipe script, it's clear what should be in place of > the broken command is something more like this: > > steps/train_mono.sh --nj 4 --cmd "$train_cmd" data/train.1k data/lang exp/mono > > Especially considering that the tutorial text refers to 'exp' but then > never shows it being used in the command: > > If the disk where you installed Kaldi is not big, you might want to make exp/ a soft link to a directory somewhere on a big disk > > On 9/25/2013 3:55 PM, Eric Riebling wrote: >> (sorry but SourceForge is having a bad hair day and refuses to >> let me log in to post to forums or bugs in a more appropriate way) >> >> There's a line in the tutorial for running: >> >> http://kaldi.sourceforge.net/tutorial_running.html >> >> that's got to be a mistake. It tells the user to type: >> >> nohup steps/train_mono.sh & >> >> But the train_mono.sh script requires arguments, and gives usage: >> >> steps/train_mono.sh >> Usage: steps/train_mono.sh [options] <data-dir> <lang-dir> <exp-dir> >> e.g.: steps/train_mono.sh data/train.1k data/lang exp/mono >> main options (for others, see top of script file) >> --config <config-file> # config containing options >> --nj <nj> # number of parallel jobs >> --feat_dim <dim> # dimension of feature vector (39) >> --cmd (utils/run.pl|utils/queue.pl <queue opts>) # how to run jobs. >> >> As a novice, it's a dead-end when something doesn't work following a tutorial, >> but I try and "forge" ahead. :) >> >> Best, > > -- > Eric Riebling Senior Research Programmer > http://ericriebling.com CMU Language Technologies Institute > 412-268-9872 GHC 6713 > > > ------------------------------------------------------------------------------ > October Webinars: Code for Performance > Free Intel webinars can help you accelerate application performance. > Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from > the latest Intel processors and coprocessors. See abstracts and register > > http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
|
From: Eric R. <er...@cs...> - 2013-09-25 19:57:47
|
From the run.sh recipe script, it's clear what should be in place of the broken command is something more like this: steps/train_mono.sh --nj 4 --cmd "$train_cmd" data/train.1k data/lang exp/mono Especially considering that the tutorial text refers to 'exp' but then never shows it being used in the command: If the disk where you installed Kaldi is not big, you might want to make exp/ a soft link to a directory somewhere on a big disk On 9/25/2013 3:55 PM, Eric Riebling wrote: > (sorry but SourceForge is having a bad hair day and refuses to > let me log in to post to forums or bugs in a more appropriate way) > > There's a line in the tutorial for running: > > http://kaldi.sourceforge.net/tutorial_running.html > > that's got to be a mistake. It tells the user to type: > > nohup steps/train_mono.sh & > > But the train_mono.sh script requires arguments, and gives usage: > > steps/train_mono.sh > Usage: steps/train_mono.sh [options] <data-dir> <lang-dir> <exp-dir> > e.g.: steps/train_mono.sh data/train.1k data/lang exp/mono > main options (for others, see top of script file) > --config <config-file> # config containing options > --nj <nj> # number of parallel jobs > --feat_dim <dim> # dimension of feature vector (39) > --cmd (utils/run.pl|utils/queue.pl <queue opts>) # how to run jobs. > > As a novice, it's a dead-end when something doesn't work following a tutorial, > but I try and "forge" ahead. :) > > Best, -- Eric Riebling Senior Research Programmer http://ericriebling.com CMU Language Technologies Institute 412-268-9872 GHC 6713 |
|
From: Eric R. <er...@cs...> - 2013-09-25 19:55:33
|
(sorry but SourceForge is having a bad hair day and refuses to let me log in to post to forums or bugs in a more appropriate way) There's a line in the tutorial for running: http://kaldi.sourceforge.net/tutorial_running.html that's got to be a mistake. It tells the user to type: nohup steps/train_mono.sh & But the train_mono.sh script requires arguments, and gives usage: steps/train_mono.sh Usage: steps/train_mono.sh [options] <data-dir> <lang-dir> <exp-dir> e.g.: steps/train_mono.sh data/train.1k data/lang exp/mono main options (for others, see top of script file) --config <config-file> # config containing options --nj <nj> # number of parallel jobs --feat_dim <dim> # dimension of feature vector (39) --cmd (utils/run.pl|utils/queue.pl <queue opts>) # how to run jobs. As a novice, it's a dead-end when something doesn't work following a tutorial, but I try and "forge" ahead. :) Best, -- Eric Riebling Senior Research Programmer http://ericriebling.com CMU Language Technologies Institute 412-268-9872 GHC 6713 |
|
From: Daniel P. <dp...@gm...> - 2013-09-23 14:51:25
|
It very much depends on what you plan to do with Kaldi. I suggest that you set up a consulting arrangement with Nickolay Shmyrev (cc'd) to figure out what it is you need to do and how to set it up. He does quite a bit of work in Israel. Since he has a lot of experience, he can set you up quite fast. Dan On Tue, Sep 17, 2013 at 6:53 AM, rafi mees <ra...@ec...> wrote: > Hello. > > my name is rafi i am part of IT staff and our development dept. requires > Kaldi. > > if you can advise us on what kind of hardware is to be used: > > cpu : dual, quad > memory: 4, 8, 16 GB > system: UBUNTU 12.10 or later. > > might i add that this would be installed on Virtual machine > > Thanks > > Rafi Mees > ECB IT Dept. > > > > > > ************************************************************************************ > This footnote confirms that this email message has been scanned by > PineApp Mail-SeCure for the presence of malicious code, vandals & computer viruses. > ************************************************************************************ > > > > > ------------------------------------------------------------------------------ > LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99! > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint > 2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack includes > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
|
From: Arnab G. <ar...@gm...> - 2013-09-23 09:45:15
|
Hi Rafi- Kaldi will run with the specs you mention. But the type of datasets you plan to use will really determine the type and amount of hardware that will be suitable. For example, if you are training/recognizing some tens of hours of speech then you may want to run at least 5-10 parallel processes, or 30-50 processes when training on a few hundred hours. For training neural networks a GPU is highly recommended (I normally use GTX 680/690). The following is from Chris Kermorvant in response to a different email: "I have configured an automatic build of Kaldi which runs the egs/rm/s5 recipe every day. It runs on a Intel(R) Xeon(R) CPU E5-2630 0 @ 2.30GHz, 12*2 CPU, 64 Gb memory. The current time for running this recipe on this computer is 55 minutes." Hope this helps. -Arnab On Tue, Sep 17, 2013 at 11:53 AM, rafi mees <ra...@ec...> wrote: > Hello. > > my name is rafi i am part of IT staff and our development dept. requires > Kaldi. > > if you can advise us on what kind of hardware is to be used: > > cpu : dual, quad > memory: 4, 8, 16 GB > system: UBUNTU 12.10 or later. > > might i add that this would be installed on Virtual machine > > Thanks > > Rafi Mees > ECB IT Dept. > > > > > > ************************************************************************************ > This footnote confirms that this email message has been scanned by > PineApp Mail-SeCure for the presence of malicious code, vandals & computer viruses. > ************************************************************************************ > > > > > ------------------------------------------------------------------------------ > LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99! > 1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint > 2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack includes > Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13. > http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
|
From: rafi m. <ra...@ec...> - 2013-09-17 11:20:19
|
Hello. my name is rafi i am part of IT staff and our development dept. requires Kaldi. if you can advise us on what kind of hardware is to be used: cpu : dual, quad memory: 4, 8, 16 GB system: UBUNTU 12.10 or later. might i add that this would be installed on Virtual machine Thanks Rafi Mees ECB IT Dept. ************************************************************************************ This footnote confirms that this email message has been scanned by PineApp Mail-SeCure for the presence of malicious code, vandals & computer viruses. ************************************************************************************ |
|
From: <fe...@in...> - 2013-09-16 16:08:11
|
Hello Chris, Thank you for your reply. Best Regards, - Felipe Espic Quoting "KERMORVANT, Christopher" <Chr...@a2...>: > Hi, > > I have configured an automatic build of Kaldi which runs the > egs/rm/s5 recipe every day. > > It runs on a Intel(R) Xeon(R) CPU E5-2630 0 @ 2.30GHz, 12*2 CPU, 64 > Gb memory. > > The current time for running this recipe on this computer is 55 minutes. > > -- > Chris > ________________________________________ > De : fe...@in... [fe...@in...] > Envoyé : jeudi 5 septembre 2013 16:25 > À : kal...@li... > Objet : [Kaldi-developers] Questions About Hardware > > Dear Sirs, > > > We are with the Speech Processing and Transmission Laboratory at > University of Chile. > We want to install Kaldi for speech recognition tasks that use the > World Street Journal Data Base (WJS0). > Link: http://www.ldc.upenn.edu/Catalog/catalogEntry.jsp?catalogId=LDC93S6A > > Currently, we don't have access to any cluster or GPU for computation. > > So, our questions are: > > 1.- Do you think it is feasible to use just a regular computer (e.g. > Intel i7, Xeon) to run Kaldi without a cluster or GPU? > > 2.- How long do you think an experiment would take for each > configuration (just one PC, GPU and cluster). I mean a rough idea > (hours, days, more than 3 days, etc)? > > > 3.- Do you suggest an economic alternative of hardware to run Kaldi > for the specified task? > > > Best Regards, > > > -Felipe Espic > > > ------------------------------------------------------------------------------ > How ServiceNow helps IT people transform IT departments: > 1. Consolidate legacy IT systems to a single system of record for IT > 2. Standardize and globalize service processes across IT > 3. Implement zero-touch automation to replace manual, redundant tasks > http://pubads.g.doubleclick.net/gampad/clk?id=51271111&iu=/4140/ostg.clktrk > _______________________________________________ > Kaldi-developers mailing list > Kal...@li... > https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
|
From: <fe...@in...> - 2013-09-16 16:07:52
|
Hello Daniel, Thank you for your reply. Best Regards, - Felipe Espic Quoting Daniel Povey <dp...@gm...>: > THe basic configuration does not require GPUs at all. You need Linux. > In a meeting-> short reply > Dan > > > On Thu, Sep 5, 2013 at 10:25 AM, <fe...@in...> wrote: >> >> Dear Sirs, >> >> >> We are with the Speech Processing and Transmission Laboratory at >> University of Chile. >> We want to install Kaldi for speech recognition tasks that use the >> World Street Journal Data Base (WJS0). >> Link: http://www.ldc.upenn.edu/Catalog/catalogEntry.jsp?catalogId=LDC93S6A >> >> Currently, we don't have access to any cluster or GPU for computation. >> >> So, our questions are: >> >> 1.- Do you think it is feasible to use just a regular computer (e.g. >> Intel i7, Xeon) to run Kaldi without a cluster or GPU? >> >> 2.- How long do you think an experiment would take for each >> configuration (just one PC, GPU and cluster). I mean a rough idea >> (hours, days, more than 3 days, etc)? >> >> >> 3.- Do you suggest an economic alternative of hardware to run Kaldi >> for the specified task? >> >> >> Best Regards, >> >> >> -Felipe Espic >> >> >> ------------------------------------------------------------------------------ >> How ServiceNow helps IT people transform IT departments: >> 1. Consolidate legacy IT systems to a single system of record for IT >> 2. Standardize and globalize service processes across IT >> 3. Implement zero-touch automation to replace manual, redundant tasks >> http://pubads.g.doubleclick.net/gampad/clk?id=51271111&iu=/4140/ostg.clktrk >> _______________________________________________ >> Kaldi-developers mailing list >> Kal...@li... >> https://lists.sourceforge.net/lists/listinfo/kaldi-developers |
23 messages has been excluded from this view by a project administrator.
