我正在运行 python 3.11 和最新版本的 llama-cpp-python 以及 一个 gguf 模型

我希望代码像聊天机器人一样正常运行，但我得到了这个错误：

traceback (most recent call last):
  file "d:ai customai arushserver.py", line 223, in <module>
    init()
  file "d:ai customai arushserver.py", line 57, in init
    m_eval(model, m_tokenize(model, prompt_init, true), false, "starting up...")
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  file "d:ai customai arushserver.py", line 182, in m_tokenize
    n_tokens = llama_cpp.llama_tokenize(
               ^^^^^^^^^^^^^^^^^^^^^^^^^
typeerror: llama_tokenize() missing 2 required positional arguments: 'add_bos' and 'special'

这是我的标记化代码：

def m_tokenize(model: llama_cpp.Llama, text: bytes, add_bos=False, special=False):
    assert model.ctx is not None
    n_ctx = llama_cpp.llama_n_ctx(model.ctx)
    tokens = (llama_cpp.llama_token * int(n_ctx))()
    n_tokens = llama_cpp.llama_tokenize(
        model.ctx,
        text,
        tokens,
        n_ctx,
        llama_cpp.c_bool(add_bos),
    )
    if int(n_tokens) < 0:
        raise RuntimeError(f'Failed to tokenize: text="{text}" n_tokens={n_tokens}')
    return list(tokens[:n_tokens])

正确答案

typeerror: llama_tokenize() missing 2 required positional arguments: 'add_bos' and 'special'

要解决该错误，您需要将参数 add_bos 和 special 包含到 llama_tokenize() 函数中。

def m_tokenize(model: llama_cpp.llama, text: bytes, add_bos=false, special=false):
    assert model.ctx is not none
    n_ctx = llama_cpp.llama_n_ctx(model.ctx)
    tokens = (llama_cpp.llama_token * int(n_ctx))()
    
    # include the missing arguments in the function call
    n_tokens = llama_cpp.llama_tokenize(
        model.ctx,
        text,
        tokens,
        n_ctx,
        # you should check if llama_cpp.c_bool(add_bos) is returning a c_boo value also you have the arguments add_bos=false and special=false in this function 
        # if i am right all you need is:
        add_bos
        # not
        # llama_cpp.c_bool(add_bos),
        # you should check if llama_cpp.c_bool(special) is returning a c_boo value
        # if i am right all you need is:
        special  # include the special argument
        # not 
        # llama_cpp.c_bool(special) 
    )
    
    if int(n_tokens) < 0:
        raise runtimeerror(f'failed to tokenize: text="{text}" n_tokens={n_tokens}')
    
    return list(tokens[:n_tokens])

来自 llama_cpp.py (github)，从 1817 开始的代码行

def llama_tokenize(
    model: llama_model_p,
    text: bytes,
    text_len: Union[c_int, int],
    tokens,  # type: Array[llama_token]
    n_max_tokens: Union[c_int, int],
    add_bos: Union[c_bool, bool],
    special: Union[c_bool, bool],
) -> int:
    """Convert the provided text into tokens."""
    return _lib.llama_tokenize(
        model, text, text_len, tokens, n_max_tokens, add_bos, special
    )


_lib.llama_tokenize.argtypes = [
    llama_model_p,
    c_char_p,
    c_int32,
    llama_token_p,
    c_int32,
    c_bool,
    c_bool,
]
_lib.llama_tokenize.restype = c_int32