Groups keyboard shortcuts have been updated

Dismiss

See shortcuts

kaldi-help

Contact owners and managers

1–30 of 8119

This is an email list for getting help on Kaldi. IMPORTANT! PLEASE GO TO http://kaldi-asr.org/forums.html TO FIND OUT HOW TO JOIN!

0 selected

QuinnQiu

12/22/24

client for online2-tcp-nnet3-decode-faster just hang-up

Hi , I use online2-tcp-nnet3-decode-faster to do online decoding, however the client just hang-up and

unread,

client for online2-tcp-nnet3-decode-faster just hang-up

Hi , I use online2-tcp-nnet3-decode-faster to do online decoding, however the client just hang-up and

12/22/24

朱志鸿, nshm...@gmail.com2

11/20/24

How to merge many asr model results?

There are many methods from old ROVER to more advanced rescoring. These days llm based merging gets

unread,

How to merge many asr model results?

There are many methods from old ROVER to more advanced rescoring. These days llm based merging gets

11/20/24

Ilya

10/15/24

ERROR with "mkgraph_lookahead.sh" SymbolTable, VectorFst, FstHeader

Good day! I installed Kaldi according to the instructions on the official website, and the rest

unread,

ERROR with "mkgraph_lookahead.sh" SymbolTable, VectorFst, FstHeader

Good day! I installed Kaldi according to the instructions on the official website, and the rest

10/15/24

Rambabu Banoth, Jan Yenda Trmal3

10/9/24

Error while running librispeech

Thank you sir, yes I have understood the problem and i am downloading again. Thank you sir On

unread,

Error while running librispeech

Thank you sir, yes I have understood the problem and i am downloading again. Thank you sir On

10/9/24

Sergei Tushev, … Daniel Povey5

10/4/24

ali to CTM

Thank you, Dan! пятница, 4 октября 2024 г. в 12:00:54 UTC+3, Daniel Povey: see get_train_ctm.sh On

unread,

ali to CTM

Thank you, Dan! пятница, 4 октября 2024 г. в 12:00:54 UTC+3, Daniel Povey: see get_train_ctm.sh On

10/4/24

Jayenthiran Pukuraj, … Abraham Nyongesa3

10/1/24

Monophone Training with TED-LIUM Dataset in Kaldi

Start with minilibrispeech. It is easier On Fri, 27 Sept 2024, 1:26 pm Jayenthiran Pukuraj, <

unread,

Monophone Training with TED-LIUM Dataset in Kaldi

Start with minilibrispeech. It is easier On Fri, 27 Sept 2024, 1:26 pm Jayenthiran Pukuraj, <

10/1/24

Jayenthiran Pukuraj

9/20/24

Kaldi Pre-trained model dataset for indian dialect

Hi im a newbie to kaldi. I have successfully installed Kaldi and have previously tested it using the

unread,

Kaldi Pre-trained model dataset for indian dialect

Hi im a newbie to kaldi. I have successfully installed Kaldi and have previously tested it using the

9/20/24

pavitra kulkarni, 10601471272

9/18/24

Noisy data issue

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

unread,

Noisy data issue

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

9/18/24

Arkadi, … Daniel Povey6

9/8/24

lattice explosion

If this is happening during training, it usually will indicate a problem, eg there was a problem with

unread,

lattice explosion

If this is happening during training, it usually will indicate a problem, eg there was a problem with

9/8/24

Sergei Tushev, Daniel Povey4

9/3/24

Mismatch between audio duration and number of frames returned by "ali-to-phones --write-lengths=true ...."

Not really, but the end effect should be fixed and symmetric so you can compensate for it in post

unread,

Mismatch between audio duration and number of frames returned by "ali-to-phones --write-lengths=true ...."

Not really, but the end effect should be fixed and symmetric so you can compensate for it in post

9/3/24

sree

8/25/24

Error while training kaldi model

I am training a speech recognition model using kaldi by ubuntu and I am getting this error when I ran

unread,

Error while training kaldi model

I am training a speech recognition model using kaldi by ubuntu and I am getting this error when I ran

8/25/24

Wizzy

7/17/24

Need help training a PLDA model

I am running a Speaker Diarization pipeline where I am experimenting with changing the embedding

unread,

Need help training a PLDA model

I am running a Speaker Diarization pipeline where I am experimenting with changing the embedding

7/17/24

Priyanshu Soni

7/17/24

Error in PLDA scoring

Hello, I fine-tuned my embedding extractor for speaker diarization on 100 hours of speaker

unread,

Error in PLDA scoring

Hello, I fine-tuned my embedding extractor for speaker diarization on 100 hours of speaker

7/17/24

Amol Bole, … Anantha Krishnan5

7/11/24

Need help on GOP

ark,t: can be used to get the result in text This code segment does the job. # Get ids for pure

unread,

Need help on GOP

ark,t: can be used to get the result in text This code segment does the job. # Get ids for pure

7/11/24

Priyanshu Soni

7/11/24

Fine-tuning the nnet3 etdnn model

I was using the pretrained etdnn model uptil now , but now I wish to fine tune it using speaker

unread,

Fine-tuning the nnet3 etdnn model

I was using the pretrained etdnn model uptil now , but now I wish to fine tune it using speaker

7/11/24

Sougata Mukherjee, … Priyanshu Soni6

7/10/24

This script is intended to be used with GPUs but you have not compiled Kaldi with CUDA If you want to use GPUs (and have them), go to src/, and configure and make on a machine where "nvcc" is installed.

I am still new at kaldi so my advice maybe downright wrong , I apologize for that in advance. But can

unread,

I am still new at kaldi so my advice maybe downright wrong , I apologize for that in advance. But can

7/10/24

Priyanshu Soni, Daniel Povey2

6/26/24

Segments file gets cleared everytime I run my code

Have a look at the wav.scp, the recording-ids should be the 1st field of wav.scp, perhaps it is

unread,

Segments file gets cleared everytime I run my code

Have a look at the wav.scp, the recording-ids should be the 1st field of wav.scp, perhaps it is

6/26/24

Anton

6/25/24

Looking for help with making a recipe for very different acoustic datasets.

I have 3 very different acoustic datasets that I'm trying to combine to train a single chain

unread,

Looking for help with making a recipe for very different acoustic datasets.

I have 3 very different acoustic datasets that I'm trying to combine to train a single chain

6/25/24

Anantha Krishnan, … Daniel Povey5

6/24/24

How to replace features in an .ark file?

sorry, copy-matrix ark:source.ark ark,t:- | less it might be called matrix-copy On Mon, Jun 24, 2024

unread,

How to replace features in an .ark file?

sorry, copy-matrix ark:source.ark ark,t:- | less it might be called matrix-copy On Mon, Jun 24, 2024

6/24/24

Foren Zas, … Daniel Povey10

6/14/24

error while training / combine the model

I think your validation set is too small and did not form at least one training example of sufficient

unread,

error while training / combine the model

I think your validation set is too small and did not form at least one training example of sufficient

6/14/24

AK Project

6/13/24

ERROR (nnet3-am-copy[5.5.1038~1-9af2c]:ExpectToken():io-funcs.cc:200) Failed to read token [started at file position -1], expected <Nnet3>

I encountered an error when executing steps/libs/nnet3/train/chain_objf/acoustic_model.py, here are

unread,

ERROR (nnet3-am-copy[5.5.1038~1-9af2c]:ExpectToken():io-funcs.cc:200) Failed to read token [started at file position -1], expected <Nnet3>

I encountered an error when executing steps/libs/nnet3/train/chain_objf/acoustic_model.py, here are

6/13/24

stilsen

6/12/24

alignments from nnet3-align-compiled only have one frame per phone

Hi, I am trying to use nnet3-align-compiled with the Gigaword XL model on a dataset I have, but the

unread,

alignments from nnet3-align-compiled only have one frame per phone

Hi, I am trying to use nnet3-align-compiled with the Gigaword XL model on a dataset I have, but the

6/12/24

Yike Zhang2

6/11/24

RTF/CPU significantly increased when more jobs used in force alignment

The RTF is 0.2 when setting nj to1, however the RTF/CPU is 0.4 when nj is set to 10. 在2024年6月11日星期二

unread,

RTF/CPU significantly increased when more jobs used in force alignment

The RTF is 0.2 when setting nj to1, however the RTF/CPU is 0.4 when nj is set to 10. 在2024年6月11日星期二

6/11/24

Dhanvi Karanam, School at Home2

5/31/24

Improper TDNN training

Hi I am part of this group. We are using Linux , an asterisk with Vosk. Kindly add me to the vosk

unread,

Improper TDNN training

Hi I am part of this group. We are using Linux , an asterisk with Vosk. Kindly add me to the vosk

5/31/24

Lucas Jo, 朱冰清2

5/30/24

Can I utilize the GrammarFst scheme in decoding graph built from CTC-topo?

I have the same problem 在2022年2月26日星期六 UTC+8 12:51:53<Lucas Jo> 写道： I am trying to use a kaldi

unread,

Can I utilize the GrammarFst scheme in decoding graph built from CTC-topo?

I have the same problem 在2022年2月26日星期六 UTC+8 12:51:53<Lucas Jo> 写道： I am trying to use a kaldi

5/30/24

Ziye Fan, … 朱冰清15

5/29/24

Q&A

Can I use grammar-fst with monophone acoustic models?

I'm a bit puzzled about left-context-phones too. Is there a way to use grammar-fst in systems

unread,

Q&A

Can I use grammar-fst with monophone acoustic models?

I'm a bit puzzled about left-context-phones too. Is there a way to use grammar-fst in systems

5/29/24

Akash Sharma

5/22/24

Vosk Speech to text

Hi, I am working on a speech to text project for converting Hindi Audio calls to transcripts. I am

unread,

Vosk Speech to text

Hi, I am working on a speech to text project for converting Hindi Audio calls to transcripts. I am

5/22/24

re...@speechmatics.com, Daniel Povey4

5/21/24

Pruning active tokens on the current frame

It would be more convenient to be able to decide to drop some tokens after decoding a frame, because

unread,

Pruning active tokens on the current frame

It would be more convenient to be able to decide to drop some tokens after decoding a frame, because

5/21/24

andrew tabarez2

5/16/24

Android armv8 build matrix lib latency

Ok after running some more tests this might be CPU / memory throttling. I run various components in

unread,

Android armv8 build matrix lib latency

Ok after running some more tests this might be CPU / memory throttling. I run various components in

5/16/24

Padala Sri Ramulu, … Anantha Krishnan3

5/14/24

How to prepare UASPEECH Data in Kaldi using SR16

https://github.com/ffxiong/uaspeech.git might be of some help. On Sunday, September 10, 2023 at 3:27:

unread,

How to prepare UASPEECH Data in Kaldi using SR16

https://github.com/ffxiong/uaspeech.git might be of some help. On Sunday, September 10, 2023 at 3:27:

5/14/24