What is the best/fast method to transcript 1K files?

DeepSpeech
#1

Hi,

I need to transcript 1K files.
What is the best/fast way to do that?
I don’t want to run 1K commands, and to spend the overhead of loading the models, etc…
Can I use the command line with more than 1 file?

Thanks,
Gil

#2

C++ tool can work on a whole directory if you pass a --audio dir/. Others don’t, but for NodeJS and Python it’s easy to write your own code as well.

#3

Thanks a lot for the fast response.

One side comment:
I just commented on the README file that the argumentes prefix (–audio, --model, etc…) are not supported in the latest deepspeech module I installed today.
From the usage text it seems that it’s still a positional arguments.

Thanks!

#4

Be aware this directory feature is just a quick hack for some testing, more than something perfect, so it’s still possible it’d be better for you to hack around in Python or JS. Regarding the arguments, you should look my answer: are you running an 0.2.0-alpha.8 build ?.