CrewAI-Studio / llms.py
zhengr's picture
init
58e0d8e
raw
history blame
3.95 kB
import os
from langchain_openai import ChatOpenAI
from langchain_groq import ChatGroq
from langchain_anthropic import ChatAnthropic
#from langchain_google_genai import ChatGoogleGenerativeAI
# from langchain_huggingface import ChatHuggingFace
from dotenv import load_dotenv
def create_openai_llm(model, temperature):
safe_pop_env_var('OPENAI_API_KEY')
safe_pop_env_var('OPENAI_API_BASE')
load_dotenv(override=True)
api_key = os.getenv('OPENAI_API_KEY')
api_base = os.getenv('OPENAI_API_BASE', 'https://api.openai.com/v1/')
if model == "gpt-4o-mini":
max_tokens = 16383
else:
max_tokens = 4095
if api_key:
return ChatOpenAI(openai_api_key=api_key, openai_api_base=api_base, model_name=model, temperature=temperature, max_tokens=max_tokens)
else:
raise ValueError("OpenAI API key not set in .env file")
def create_anthropic_llm(model, temperature):
api_key = os.getenv('ANTHROPIC_API_KEY')
if api_key:
return ChatAnthropic(anthropic_api_key=api_key, model_name=model, temperature=temperature,max_tokens=4095)
else:
raise ValueError("Anthropic API key not set in .env file")
def create_groq_llm(model, temperature):
api_key = os.getenv('GROQ_API_KEY')
if api_key:
return ChatGroq(groq_api_key=api_key, model_name=model, temperature=temperature, max_tokens=4095)
else:
raise ValueError("Groq API key not set in .env file")
# def create_googleai_llm(model, temperature):
# api_key = os.getenv('GOOGLE_API_KEY')
# if api_key:
# return ChatGoogleGenerativeAI(model=model, temperature=temperature)
# else:
# raise ValueError("Google AI API key not set in .env file")
# def create_huggingfacehub_llm(model, temperature):
# api_key = os.getenv('HUGGINGFACEHUB_API_KEY')
# if api_key:
# return ChatHuggingFace(repo_id=model, huggingfacehub_api_token=api_key, model_kwargs={"temperature":temperature, "max_tokens": 4096})
# else:
# raise ValueError("HuggingFace API key not set in .env file")
def create_lmstudio_llm(model, temperature):
api_base = os.getenv('LMSTUDIO_API_BASE')
os.environ["OPENAI_API_KEY"] = "lm-studio"
os.environ["OPENAI_API_BASE"] = api_base
if api_base:
return ChatOpenAI(openai_api_key='lm-studio', openai_api_base=api_base, temperature=temperature, max_tokens=4095)
else:
raise ValueError("LM Studio API base not set in .env file")
LLM_CONFIG = {
"OpenAI": {
"models": ["gpt-4o","gpt-4o-mini","gpt-3.5-turbo", "gpt-4-turbo"],
"create_llm": create_openai_llm
},
"Groq": {
"models": ["llama3-8b-8192","llama3-70b-8192", "mixtral-8x7b-32768"],
"create_llm": create_groq_llm
},
# "GoogleAI": {
# "models": ["gemini-1.5-pro","gemini-1.5-flash", "gemini-1.0-pro"],
# "create_llm": create_googleai_llm
# },
# "HuggingFaceHub": {
# "models": ["mistralai/Mistral-7B-Instruct-v0.2", "mistralai/Codestral-22B-v0.1", "EleutherAI/gpt-neo-2.7B"],
# "create_llm": create_huggingfacehub_llm
# },
"Anthropic": {
"models": ["claude-3-5-sonnet-20240620"],
"create_llm": create_anthropic_llm
},
"LM Studio": {
"models": ["lms-default"],
"create_llm": create_lmstudio_llm
}
}
def llm_providers_and_models():
return [f"{provider}: {model}" for provider in LLM_CONFIG.keys() for model in LLM_CONFIG[provider]["models"]]
def create_llm(provider_and_model, temperature=0.1):
provider, model = provider_and_model.split(": ")
create_llm_func = LLM_CONFIG.get(provider, {}).get("create_llm")
if create_llm_func:
return create_llm_func(model, temperature)
else:
raise ValueError(f"LLM provider {provider} is not recognized or not supported")
def safe_pop_env_var(key):
try:
os.environ.pop(key)
except KeyError:
pass