By "duplicate file" I'm guessing you mean files with identical content, rather than files with the same name but in different subdirectories??

In the former case, read in each file, calculate an MD5 checksum of its contents, and use that as a key to a hash, whose values are the pathnames.

Dave.