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

# API_URL = "/static-proxy?url=https%3A%2F%2Fapi-inference.huggingface.co%2Fmodels%2Fmistralai%2FMistral-7B-v0.1%26quot%3B%3C%2Fspan%3E%3C!-- HTML_TAG_END -->
# headers = {"Authorization": f"Bearer {key}"}

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

# def LLM(llm_name, length):
#     print(llm_name)
#     tokenizer = AutoTokenizer.from_pretrained(llm_name)
#     model = AutoModelForCausalLM.from_pretrained(llm_name,
#                                                  trust_remote_code=True, 
#                                                  device_map="auto",
#                                                  load_in_8bit=True)
#     pipe = pipeline("text-generation",
#                     model=model,
#                     tokenizer=tokenizer,
#                     max_length=length,
#                     do_sample=True,
#                     top_p=0.95,
#                     repetition_penalty=1.2,
#                    )
#     return pipe
tokenizer = AutoTokenizer.from_pretrained('replit/replit-code-v1_5-3b', trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained('replit/replit-code-v1_5-3b', trust_remote_code=True)

x = tokenizer.encode('def fibonacci(n): ', return_tensors='pt')
y = model.generate(x, max_length=100, do_sample=True, top_p=0.95, top_k=4, temperature=0.2, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id)

# decoding
generated_code = tokenizer.decode(y[0], skip_special_tokens=True, clean_up_tokenization_spaces=False)
print(generated_code)

pipe = generated_code
# tokenizer = AutoTokenizer.from_pretrained("WizardLM/WizardCoder-1B-V1.0")
# base_model = AutoModelForCausalLM.from_pretrained("WizardLM/WizardCoder-1B-V1.0")
# Mistral 7B
# mistral_llm = LLM("mistralai/Mistral-7B-v0.1",30000)
mistral_llm = pipe

# WizardCoder 13B
# wizard_llm = LLM("WizardLM/WizardCoder-Python-13B-V1.0",8000)
wizard_llm = pipe
# hf_llm = HuggingFacePipeline(pipeline=pipe)

def ask_model(model, prompt):
    if(model == 'mistral'):
        return mistral_llm(prompt)
    if(model == 'wizard'):
        return wizard_llm(prompt)






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,
)


@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}








@app.post("/ask_llm")
async def ask_llm_endpoint(model:str, prompt: str):
    result = ask_model(model,prompt)
    return {"result": result}









# APIs

# @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}