vercel llm api
1.0.0
這是Vercel AI遊樂場的逆向工程API包裝器,它允許免費訪問許多LLM,包括Openai的Chatgpt,Cohere's Commery Nightly以及一些開源型號。
用markdown-toc生成的目錄。
您可以通過運行以下命令來安裝此庫:
pip3 install vercel-llm-api
示例可以在/examples目錄中找到。要運行這些示例,只需從終端執行隨附的Python文件即可。
python3 examples/generate.py
要使用此庫,只需導入vercel_ai並創建一個vercel_ai.Client實例。您可以使用proxy關鍵字參數指定代理。
正常示例:
import vercel_ai
client = vercel_ai . Client ()代理示例:
import vercel_ai
client = vercel_ai . Client ( proxy = "socks5h://193.29.62.48:11003" )請注意,以下示例假設client是您的vercel_ai.Client實例的名稱。
客戶在初始化時下載可用模型,並將其存儲在client.models中。
>> > print ( json . dumps ( client . models , indent = 2 ))
{
"anthropic:claude-instant-v1" : {
"id" : "anthropic:claude-instant-v1" , #the model's id
"provider" : "anthropic" , #the model's provider
"providerHumanName" : "Anthropic" , #the provider's display name
"makerHumanName" : "Anthropic" , #the maker of the model
"minBillingTier" : "hobby" , #the minimum billing tier needed to use the model
"parameters" : { #a dict of optional parameters that can be passed to the generate function
"temperature" : { #the name of the parameter
"value" : 1 , #the default value for the parameter
"range" : [ 0 , 1 ] #a range of possible values for the parameter
},
...
}
...
}
}請注意,由於尚無AUTH,因此如果模型具有"minBillingTier"屬性,則無法使用它。
client.model_ids中也可用模型ID列表。
>> > print ( json . dumps ( client . model_ids , indent = 2 ))
[
"anthropic:claude-instant-v1" , #locked to hobby tier; unusable
"anthropic:claude-v1" , #locked to hobby tier; unusable
"replicate:replicate/alpaca-7b" ,
"replicate:stability-ai/stablelm-tuned-alpha-7b" ,
"huggingface:bigscience/bloom" ,
"huggingface:bigscience/bloomz" ,
"huggingface:google/flan-t5-xxl" ,
"huggingface:google/flan-ul2" ,
"huggingface:EleutherAI/gpt-neox-20b" ,
"huggingface:OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5" ,
"huggingface:bigcode/santacoder" ,
"cohere:command-medium-nightly" ,
"cohere:command-xlarge-nightly" ,
"openai:gpt-4" , #locked to pro tier; unusable
"openai:code-cushman-001" ,
"openai:code-davinci-002" ,
"openai:gpt-3.5-turbo" ,
"openai:text-ada-001" ,
"openai:text-babbage-001" ,
"openai:text-curie-001" ,
"openai:text-davinci-002" ,
"openai:text-davinci-003"
]可以在client.model_params上找到每個模型的默認參數的命令。
>> > print ( json . dumps ( client . model_defaults , indent = 2 ))
{
"anthropic:claude-instant-v1" : {
"temperature" : 1 ,
"maximumLength" : 200 ,
"topP" : 1 ,
"topK" : 1 ,
"presencePenalty" : 1 ,
"frequencyPenalty" : 1 ,
"stopSequences" : [
" n n Human:"
]
},
...
}要生成一些文本,請使用client.generate函數,該函數接受以下參數:
model - 您要使用的模型的ID。prompt - 您的提示。params = {} - 可選參數的dist。有關如何找到這些信息,請參見上一節。該函數是一個發電機,將新生成的文本返回為字符串。
流示例:
for chunk in client . generate ( "openai:gpt-3.5-turbo" , "Summarize the GNU GPL v3" ):
print ( chunk , end = "" , flush = True )非流程示例:
result = ""
for chunk in client . generate ( "openai:gpt-3.5-turbo" , "Summarize the GNU GPL v3" ):
result += chunk
print ( result )要生成聊天消息,請使用client.chat函數,該函數接受以下參數:
model - 您要使用的模型的ID。messages - 消息列表。這種格式與您如何使用官方OpenAI API相同。params = {} - 可選參數的dist。有關如何找到這些內容,請參見“下載可用模型”部分。該函數是一個發電機,將新生成的文本返回為字符串。
messages = [
{ "role" : "system" , "content" : "You are a helpful assistant." },
{ "role" : "user" , "content" : "Who won the world series in 2020?" },
{ "role" : "assistant" , "content" : "The Los Angeles Dodgers won the World Series in 2020." },
{ "role" : "user" , "content" : "Where was it played?" }
]
for chunk in client . chat ( "openai:gpt-3.5-turbo" , messages ):
print ( chunk , end = "" , flush = True )
print ()如果要顯示調試消息,只需致電vercel_ai.logger.setLevel 。
import vercel_ai
import logging
vercel_ai . logger . setLevel ( logging . INFO )該程序在GNU GPL V3下獲得許可。所有代碼都是由我撰寫的,Ading2210。
ading2210/vercel-llm-api: a reverse engineered API wrapper for the Vercel AI Playground
Copyright (C) 2023 ading2210
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <https://www.gnu.org/licenses/>.