github/chatgpt-on-wechat
1
0
Fork 0
mirror of https://github.com/zhayujie/chatgpt-on-wechat.git synced 2026-03-28 05:58:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: update counting tokens for new models

Browse Source
This commit is contained in:
lanvent
2023-06-14 13:36:07 +08:00
parent a582a46ce9
commit bb737a71d5
2 changed files with 9 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
bot/chatgpt/chat_gpt_bot.py
View File

@@ -121,6 +121,7 @@ class ChatGPTBot(Bot, OpenAIImage):
if args is None: if args is None:
args = self.args args = self.args
response = openai.ChatCompletion.create(api_key=api_key, messages=session.messages, **args) response = openai.ChatCompletion.create(api_key=api_key, messages=session.messages, **args)
# logger.debug("[CHATGPT] response={}".format(response))
# logger.info("[ChatGPT] reply={}, total_tokens={}".format(response.choices[0]['message']['content'], response["usage"]["total_tokens"])) # logger.info("[ChatGPT] reply={}, total_tokens={}".format(response.choices[0]['message']['content'], response["usage"]["total_tokens"]))
return { return {
"total_tokens": response["usage"]["total_tokens"], "total_tokens": response["usage"]["total_tokens"],

16
bot/chatgpt/chat_gpt_session.py
View File

@@ -57,25 +57,25 @@ def num_tokens_from_messages(messages, model):
"""Returns the number of tokens used by a list of messages.""" """Returns the number of tokens used by a list of messages."""
import tiktoken import tiktoken
if model == "gpt-3.5-turbo" or model == "gpt-35-turbo": if model in ["gpt-3.5-turbo-0301", "gpt-3.5-turbo-0613", "gpt-35-turbo", "gpt-3.5-turbo-16k", "gpt-3.5-turbo-16k-0613"]:
return num_tokens_from_messages(messages, model="gpt-3.5-turbo-0301") return num_tokens_from_messages(messages, model="gpt-3.5-turbo")
elif model == "gpt-4": elif model in ["gpt-4-0314", "gpt-4-0613", "gpt-4-32k", "gpt-4-32k-0613"]:
return num_tokens_from_messages(messages, model="gpt-4-0314") return num_tokens_from_messages(messages, model="gpt-4")
try: try:
encoding = tiktoken.encoding_for_model(model) encoding = tiktoken.encoding_for_model(model)
except KeyError: except KeyError:
logger.debug("Warning: model not found. Using cl100k_base encoding.") logger.debug("Warning: model not found. Using cl100k_base encoding.")
encoding = tiktoken.get_encoding("cl100k_base") encoding = tiktoken.get_encoding("cl100k_base")
if model == "gpt-3.5-turbo-0301": if model == "gpt-3.5-turbo":
tokens_per_message = 4 # every message follows <|start|>{role/name}\n{content}<|end|>\n 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 tokens_per_name = -1 # if there's a name, the role is omitted
elif model == "gpt-4-0314": elif model == "gpt-4":
tokens_per_message = 3 tokens_per_message = 3
tokens_per_name = 1 tokens_per_name = 1
else: else:
logger.warn(f"num_tokens_from_messages() is not implemented for model {model}. Returning num tokens assuming gpt-3.5-turbo-0301.") logger.warn(f"num_tokens_from_messages() is not implemented for model {model}. Returning num tokens assuming gpt-3.5-turbo.")
return num_tokens_from_messages(messages, model="gpt-3.5-turbo-0301") return num_tokens_from_messages(messages, model="gpt-3.5-turbo")
num_tokens = 0 num_tokens = 0
for message in messages: for message in messages:
num_tokens += tokens_per_message num_tokens += tokens_per_message