There's probably a way to do this while going through the data only once, but I can't think of it. Here's a method that goes through the values (file sizes) once and then through the keys. It saves the number of times each file size is seen, and then prints out the files that have sizes that were seen more than once.

#!/usr/bin/perl

my %sizes = qw( a 1 b 2 c 3 d 3 e 2 f 2);
my %times_seen;

for my $v (values %sizes){
    $times_seen{$v} ? $times_seen{$v}++ : ($times_seen{$v} = 1);
}
for my $k (keys %sizes){
    if($times_seen{$sizes{$k}} > 1 ){
        print "$k => $times_seen{$sizes{$k}}\n";
    }
}
[download]