Now that your expected output looks more logical, a simple test can illustrate one possible solution.

use strict;
use warnings;

use Test::More;

my @data = (
    {
        have => '/fd/gfree/tere/frf4545/geerg/fds/0.1/fsdf/dsakdsa/',
        want => '/fd/gfree/tere/frf4545/geerg/fds/0.1',
    },
    {
        have => '/fd/gfree/tere/frf4545/geerg/dfds/5.9/fdsf/fdsfd/',
        want => '/fd/gfree/tere/frf4545/geerg/dfds/5.9',
    },
    {
        have => '/fd/gfree/tere/frf4545/geerg/dsad/02.44/fdsf/fdsf/',
        want => '/fd/gfree/tere/frf4545/geerg/dsad/02.44',
    },
);

plan tests => scalar @data;

for my $t (@data) {
    $t->{have} =~ s/(\.\d+).*?$/$1/;
    is $t->{have}, $t->{want};
}
[download]

Perfect! Thanks for your time. Your solution works perfectly for me.

If you look at the code I posted you see that I have a couple of / in it in addition to the regex I mention in the text. So the code really looks for a sequence of digits and dots between two slashes. It is true it would truncate the path after the first occurrence of such a pattern.