Why does the token vocabs are unreadable code? Is there any special handling when training BPE by sentencepiece?
· Sign up or log in to comment
