def get_nonascii_toks()

in pyrit/auxiliary_attacks/gcg/attack/base/attack_manager.py [0:0]

• 16 lines of code
• 8 McCabe index (conditional complexity)

def get_nonascii_toks(tokenizer, device="cpu"):

    def is_ascii(s):
        return s.isascii() and s.isprintable()

    ascii_toks = []
    for i in range(3, tokenizer.vocab_size):
        if not is_ascii(tokenizer.decode([i])):
            ascii_toks.append(i)

    if tokenizer.bos_token_id is not None:
        ascii_toks.append(tokenizer.bos_token_id)
    if tokenizer.eos_token_id is not None:
        ascii_toks.append(tokenizer.eos_token_id)
    if tokenizer.pad_token_id is not None:
        ascii_toks.append(tokenizer.pad_token_id)
    if tokenizer.unk_token_id is not None:
        ascii_toks.append(tokenizer.unk_token_id)

    return torch.tensor(ascii_toks, device=device)