ver0.4.6 0613接口更新、支持azure小变动

chatgpt_tool_hub/common/calculate_token.py

@@ -2,29 +2,42 @@
 
 
 # refer to https://github.com/openai/openai-cookbook/blob/main/examples/How_to_count_tokens_with_tiktoken.ipynb
-def count_message_tokens(messages, model: str = "gpt-3.5-turbo"):
-    """Returns the number of tokens used by a list of messages."""
+
+def count_message_tokens(messages, model="gpt-3.5-turbo-0613"):
     import tiktoken
+
+    if model.startswith("gpt-35"):
+        model = model.replace("gpt-35-", "gpt-3.5-")
+
+    """Return the number of tokens used by a list of messages."""
     try:
         encoding = tiktoken.encoding_for_model(model)
     except KeyError:
         LOG.debug("Warning: model not found. Using cl100k_base encoding.")
         encoding = tiktoken.get_encoding("cl100k_base")
-    if model != "gpt-3.5-turbo-0613" and model.startswith("gpt-3.5-turbo"):
-        return count_message_tokens(messages, model="gpt-3.5-turbo-0613")
-    elif model == "gpt-4":
-        return count_message_tokens(messages, model="gpt-4-0314")
-    elif model == "gpt-3.5-turbo-0613":
-        # 注意，0613 3.5版本特殊token计算数默认与0301一致，但未经测试
-        tokens_per_message = 4  # every message follows <|start|>{role/name}\n{content}<|end|>\n
-        tokens_per_name = -1  # if there's a name, the role is omitted
-    elif model == "gpt-4-0314":
+    if model in {
+        "gpt-3.5-turbo-0613",
+        "gpt-3.5-turbo-16k-0613",
+        "gpt-4-0314",
+        "gpt-4-32k-0314",
+        "gpt-4-0613",
+        "gpt-4-32k-0613",
+        }:
         tokens_per_message = 3
         tokens_per_name = 1
-    else:
-        LOG.warn(f"count_message_tokens() is not implemented for model {model}. "
-                 f"Returning num tokens assuming gpt-3.5-turbo-0613.")
+    elif model == "gpt-3.5-turbo-0301":
+        tokens_per_message = 4  # every message follows <|start|>{role/name}\n{content}<|end|>\n
+        tokens_per_name = -1  # if there's a name, the role is omitted
+    elif "gpt-3.5-turbo" in model:
+        LOG.debug("Warning: gpt-3.5-turbo may update over time. Returning num tokens assuming gpt-3.5-turbo-0613.")
         return count_message_tokens(messages, model="gpt-3.5-turbo-0613")
+    elif "gpt-4" in model:
+        LOG.debug("Warning: gpt-4 may update over time. Returning num tokens assuming gpt-4-0613.")
+        return count_message_tokens(messages, model="gpt-4-0613")
+    else:
+        raise NotImplementedError(
+            f"""num_tokens_from_messages() is not implemented for model {model}. See https://github.com/openai/openai-python/blob/main/chatml.md for information on how messages are converted to tokens."""
+        )
     num_tokens = 0
     for message in messages:
         num_tokens += tokens_per_message
@@ -36,17 +49,21 @@ def count_message_tokens(messages, model: str = "gpt-3.5-turbo"):
     return num_tokens
 
 
-def count_string_tokens(string: str, model_name: str = "gpt-3.5-turbo") -> int:
+def count_string_tokens(string: str, model_name: str = "gpt-35-turbo") -> int:
     """
     Returns the number of tokens in a text string.
 
     Args:
     string (str): The text string.
-    model_name (str): The name of the encoding to use. (e.g., "gpt-3.5-turbo")
+    model_name (str): The name of the encoding to use. (e.g., "gpt-35-turbo")
 
     Returns:
     int: The number of tokens in the text string.
     """
     import tiktoken
+
+    if model_name.startswith("gpt-35"):
+        model_name = model_name.replace("gpt-35-", "gpt-3.5-")
+
     encoding = tiktoken.encoding_for_model(model_name)
     return len(encoding.encode(string))

chatgpt_tool_hub/models/__init__.py

@@ -8,7 +8,7 @@
 verbose: bool = False
 llm_cache: Optional[BaseCache] = None
 
-# default for gpt-3.5
+# default for gpt-35
 ALL_MAX_TOKENS_NUM = 4000
 
 # token manage strategy
@@ -26,7 +26,7 @@ def change_memory_max_tokens(memory_max_tokens_num: int):
 def build_model_params(kwargs: dict) -> dict:
     _api_key = get_from_dict_or_env(kwargs, "llm_api_key", "LLM_API_KEY")
     _proxy = get_from_dict_or_env(kwargs, "proxy", "PROXY", "")
-    _model = get_from_dict_or_env(kwargs, "model_name", "MODEL_NAME", "gpt-3.5-turbo")
+    _model = get_from_dict_or_env(kwargs, "model_name", "MODEL_NAME", "gpt-35-turbo")
     _timeout = get_from_dict_or_env(kwargs, "request_timeout", "REQUEST_TIMEOUT", 120)
     _llm_api_base_url = get_from_dict_or_env(kwargs, "llm_api_base_url", "LLM_API_BASE_URL", openai_default_api_base)
     _deployment_id = get_from_dict_or_env(kwargs, "deployment_id", "DEPLOYMENT_ID", "")

chatgpt_tool_hub/models/chatgpt/chatgpt.py

@@ -114,12 +114,12 @@ class ChatOpenAI(BaseChatModel, BaseModel):
         .. code-block:: python
 
             from lib.chat_models import ChatOpenAI
-            openai = ChatOpenAI(model_name="gpt-3.5-turbo")
+            openai = ChatOpenAI(model_name="gpt-35-turbo")
     """
 
     client: Any  #: :meta private:
     """Model name to use."""
-    model_name: str = "gpt-3.5-turbo"
+    model_name: str = "gpt-35-turbo"
     """Holds any model parameters valid for `create` call not explicitly specified."""
     model_kwargs: Dict[str, Any] = Field(default_factory=dict)
     """llm api base url"""
@@ -367,19 +367,19 @@ def get_num_tokens(self, text: str) -> int:
                 "This is needed in order to calculate get_num_tokens. "
                 "Please it install it with `pip install tiktoken`."
             )
-        # create a GPT-3.5-Turbo encoder instance
+        # create a gpt-35-Turbo encoder instance
         enc = tiktoken.encoding_for_model(self.model_name)
 
-        # encode the text using the GPT-3.5-Turbo encoder
+        # encode the text using the gpt-35-Turbo encoder
         tokenized_text = enc.encode(text)
 
         # calculate the number of tokens in the encoded text
         return len(tokenized_text)
 
     def get_num_tokens_from_messages(
-        self, messages: List[BaseMessage], model: str = "gpt-3.5-turbo"
+        self, messages: List[BaseMessage], model: str = "gpt-35-turbo"
     ) -> int:
-        """Calculate num tokens for gpt-3.5-turbo with tiktoken package."""
+        """Calculate num tokens for gpt-35-turbo with tiktoken package."""
         try:
             import tiktoken
         except ImportError:
@@ -394,7 +394,7 @@ def get_num_tokens_from_messages(
             encoding = tiktoken.encoding_for_model(model)
         except KeyError:
             encoding = tiktoken.get_encoding("cl100k_base")
-        if not model.startswith("gpt-3.5-turbo"):
+        if not model.startswith("gpt-35-turbo"):
             raise NotImplementedError(
                 f"get_num_tokens_from_messages() is not presently implemented "
                 f"for model {model}."

chatgpt_tool_hub/models/model_factory.py

@@ -21,7 +21,7 @@ def match_model(self, name: str) -> str:
         return ""
 
     def create_llm_model(self, **model_kwargs):
-        _model = get_from_dict_or_env(model_kwargs, "model_name", "MODEL_NAME", "gpt-3.5-turbo")
+        _model = get_from_dict_or_env(model_kwargs, "model_name", "MODEL_NAME", "gpt-35-turbo")
         match_llm_model = self.match_model(_model)
         if match_llm_model == "chatgpt":
             return ChatOpenAI(**model_kwargs)

chatgpt_tool_hub/version.py

@@ -1 +1 @@
-__version__ = '0.4.5'
+__version__ = '0.4.6'

requirements.txt

@@ -21,8 +21,8 @@ beautifulsoup4~=4.12.0
 
 # for openai
 tenacity~=8.2.2
-openai~=0.27.2
-tiktoken~=0.3.2
+openai~=0.27.4
+tiktoken~=0.4.0
 
 # all tool dependencies
 arxiv

requirements_all.txt

@@ -24,11 +24,11 @@ pyopenssl
 # for models openai
 tenacity~=8.2.2
 
-openai~=0.27.2
+openai~=0.27.4
 
 SQLAlchemy~=2.0.7
 
-tiktoken~=0.3.2
+tiktoken~=0.4.0
 
 arxiv
 

setup.py

@@ -36,8 +36,8 @@
         'lxml',
         'beautifulsoup4~=4.12.0',
         "tenacity~=8.2.2",
-        "openai~=0.27.2",
-        "tiktoken~=0.3.2",
+        "openai~=0.27.4",
+        "tiktoken~=0.4.0",
         "arxiv",
         "wikipedia",
         "wolframalpha",

terminal_io.py

@@ -122,7 +122,7 @@ def __init__(self, timeout: int):
 
         self.app = self.create_app()
 
-        self.model = 'gpt-3.5-turbo'
+        self.model = 'gpt-35-turbo'
         self.timeout = timeout
 
         # todo 需要llm-os model支持
@@ -215,7 +215,7 @@ def set_model(self, new_model: str):
             tokens_limit = 32768
         elif self.model.startswith("gpt-4"):
             tokens_limit = 8192
-        elif self.model.startswith("gpt-3.5-turbo"):
+        elif self.model.startswith("gpt-35-turbo"):
             tokens_limit = 4096
         else:
             console.print(f"[red]没有该模型 {new_model} tokens信息，模型未变更: [bold cyan]{old_model}[/].")
@@ -246,11 +246,11 @@ class CustomCompleter(Completer):
     ]
 
     available_models = [
-        "gpt-3.5-turbo",
-        "gpt-3.5-turbo-0301",
-        "gpt-3.5-turbo-0613",
-        "gpt-3.5-turbo-16k",
-        "gpt-3.5-turbo-16k-0613",
+        "gpt-35-turbo",
+        "gpt-35-turbo-0301",
+        "gpt-35-turbo-0613",
+        "gpt-35-turbo-16k",
+        "gpt-35-turbo-16k-0613",
         "gpt-4",
         "gpt-4-0314",
         "gpt-4-32k",
@@ -478,7 +478,7 @@ def handle_command(command: str, llm_os: LLMOS):
     /depth                   - 设置LLM-OS思考深度 (设置过大可能无法停止)
     /reset                   - LLM-OS重置 (重新加载配置并重置聊天记录)
     /timeout \[new_timeout]   - 修改访问llm的请求超时时间
-    /model   \[model_name]    - 切换模型 (目前仅支持gpt-3.5)
+    /model   \[model_name]    - 切换模型 (目前仅支持gpt-35)
     /last                    - 显示上一次LLM-OS的回复内容
     /copy                    - 复制上一次LLM-OS的回复内容到粘贴板
     /undo                    - 清除上一次与llm的对话记录 (包含问题和回复)