Another possible solution for you...Since no array of the all the "line words" are created, this scales linearly with file size.

#!/usr/bin/perl
use warnings;
use strict;

my @tokens;
my $n=2; #adjust as needed

while (my $line = <DATA>)
{
   chomp $line;
   push @tokens, $line;
   next if @tokens <$n;
   
   print "@tokens\n";
   shift @tokens;
}

=prints:
N=2....
Hello I
I am
am happy
happy this
this is
is the
the text
text to
to play
play with

N=3....
Hello I am
I am happy
am happy this
happy this is
this is the
is the text
the text to
text to play
to play with
=cut

__DATA__
Hello
I
am
happy
this
is
the
text
to
play
with
[download]