This blog is about running pre-trained Spanish model using Montreal Force Aligner.
In this tutorial I will be using Ubuntu 16.04lts.Let’s get started.
Compile Kaldi :- More information about compiling Kaldi is given in tools and src directory
under the green icon named “INSTALL”.
First follow instruction in tools section and then under src section.
First step is to check dependencies.
Command > extras/check_dependencies.sh
Such message should be printed out.This script checks extra dependencies needed for kaldi.
If your system default C++ compiler is not supported, you can do the check with another
by setting the CXX environment variable, e.g.
Command > make
This command installs various libraries to support kaldi features such as ATLAS,openfst,SCTK,
sph2pipe and CLAPACK headers.
If your system default compiler does not have adequate support for C++11, you can specify a
C+11 compliant compiler as a command argument, e.g.
If you have multiple CPUs and want to speed things up, you can do a parallel build by supplying
the "-j" option to make, e.g. to use 4 CPUs
make -j 4
Look you should be in tools directory in kaldi.
Now change your working directory to src
The installation instructions are
make depend -j 8
make -j 8
Note that we added the "-j 8" to run in parallel because "make" takes a long
time. 8 jobs might be too many for a laptop or small desktop machine with not
For any further information visit http://kaldi-asr.org/doc/
and head straight to "The build process (how Kaldi is compiled)".
With all this processes done.We have completed process of compiling Kaldi.
We would align using pre-trained models available on Montreal Force Aligner
Download source code from github.