Re: [Kaldi-developers] Please add scripts for checking input data

[Bo T. <dx...@16...>]

Thank you!
I'm sorry for not checking the existing scripts but KALDI is too powerful to work through :)

于 8/5/13 1:19 AM, Daniel Povey 写道:

You may also try the script utils/fix_data_dir.sh which automatically
fixes issues like sorting, IIRC.
Dan


On Sun, Aug 4, 2013 at 1:16 PM, Daniel Povey <dp...@gm...> wrote:

This has been recently added-- you are using the trunk, right?  It's
utils/validate_data_dir.sh
Dan


On Fri, Aug 2, 2013 at 8:14 AM, Bo Tian <dx...@16...> wrote:

Hello everyone,

I love KALDI very much for its variety of abilities to build models and
decode.

But could you please add an script to check the format of the input data
for training?

Today I tried to build acoustic model using my own data, I found it kept
failing because of unsorted text(spk2utt, etc.), and I sorted them
myself but it didn't work. Finally, I found it was advised to append
speaker id to utterance id for sorting.

Such problem is not difficult to solve, but it's difficult to check if
it's solved. Because I need to extract feat again and again(maybe it's
not needed but I don't know), and after run train_mon.sh for a while, it
tells me the error. It takes too tong. I think all these could be done
after I finish preparing the data before training phase.

I will appreciate your help!

Sincerely

--
Department of Computer Science and Technology, Tsinghua University
Bo Tian



------------------------------------------------------------------------------
Get your SQL database under version control now!
Version control is standard for application code, but databases havent
caught up. So what steps can you take to put your SQL databases under
version control? Why should you start doing it? Read more to find out.
http://pubads.g.doubleclick.net/gampad/clk?id=49501711&iu=/4140/ostg.clktrk
_______________________________________________
Kaldi-developers mailing list
Kal...@li...https://lists.sourceforge.net/lists/listinfo/kaldi-developers



-- 
Department of Computer Science and Technology, Tsinghua University
Bo Tian