endpoints.py

from fastapi import FastAPI
from fastapi.middleware.cors import CORSMiddleware
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
import torch
import os
import requests
# from langchain.llms.huggingface_pipeline import HuggingFacePipeline

key = os.environ.get("huggingface_key")
openai_api_key = os.environ.get("openai_key")
app = FastAPI(openapi_url="/api/v1/LLM/openapi.json", docs_url="/api/v1/LLM/docs")

app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_methods=["*"],
    allow_headers=["*"],
    allow_credentials=True,
)
API_URL = "https://api-inference.huggingface.co/models/mistralai/Mistral-7B-v0.1"
headers = {"Authorization": f"Bearer {key}"}

def query(payload):
	response = requests.post(API_URL, headers=headers, json=payload)
	return response.json()







# tokenizer = AutoTokenizer.from_pretrained("WizardLM/WizardCoder-1B-V1.0")
# base_model = AutoModelForCausalLM.from_pretrained("WizardLM/WizardCoder-1B-V1.0")
model = "mistralai/Mixtral-8x7B-Instruct-v0.1"
tokenizer = AutoTokenizer.from_pretrained(model)
base_model = AutoModelForCausalLM.from_pretrained(model)
pipe = pipeline("text-generation",
                model=base_model,
                tokenizer=tokenizer,
                max_length=4000,
                do_sample=True,
                top_p=0.95,
                repetition_penalty=1.2,
               )
# hf_llm = HuggingFacePipeline(pipeline=pipe)


@app.get("/")
def root():
    return {"message": "R&D LLM API"}
@app.get("/get")
def get():
    result = pipe("name 5 programming languages",do_sample=False)
    print(result)
    return {"message": result}


async def askLLM(prompt):
    output = pipe(prompt,do_sample=False)
    return output

@app.post("/ask_llm")
async def ask_llm_endpoint(prompt: str):
    # result = await askLLM(prompt)
    result = pipe(prompt,do_sample=False)
    return {"result": result}


@app.post("/ask_HFAPI")
def ask_HFAPI_endpoint(prompt: str):
    result = query(prompt)
    return {"result": result}
    
from langchain.llms import OpenAI

llm = OpenAI(model_name="text-davinci-003", temperature=0.5, openai_api_key=openai_api_key)

@app.post("/ask_GPT")
def ask_GPT_endpoint(prompt: str):
    result = llm(prompt)
    return {"result": result}