app.py

from fastapi import FastAPI, Request
from fastapi.responses import HTMLResponse, JSONResponse
import uvicorn
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, TextIteratorStreamer
from threading import Thread

app = FastAPI()

# Chargement du modèle uniquement si CUDA est disponible
if torch.cuda.is_available():
    model_id = "mistralai/Mistral-7B-Instruct-v0.3"
    model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map="auto")
    tokenizer = AutoTokenizer.from_pretrained(model_id)
else:
    model = None
    tokenizer = None

MAX_INPUT_TOKEN_LENGTH = 4096

def generate_response(message: str, history: list) -> str:
    conversation = history + [{"role": "user", "content": message}]
    input_ids = tokenizer.apply_chat_template(conversation, return_tensors="pt")
    if input_ids.shape[1] > MAX_INPUT_TOKEN_LENGTH:
        input_ids = input_ids[:, -MAX_INPUT_TOKEN_LENGTH:]
    input_ids = input_ids.to(model.device)
    
    streamer = TextIteratorStreamer(tokenizer, timeout=20.0, skip_prompt=True, skip_special_tokens=True)
    generate_kwargs = {
        "input_ids": input_ids,
        "streamer": streamer,
        "max_new_tokens": 1024,
        "do_sample": True,
        "top_p": 0.9,
        "top_k": 50,
        "temperature": 0.6,
        "num_beams": 1,
        "repetition_penalty": 1.2,
    }
    t = Thread(target=model.generate, kwargs=generate_kwargs)
    t.start()
    
    response_text = ""
    for text in streamer:
        response_text += text
    return response_text

@app.post("/chat")
async def chat_endpoint(request: Request):
    data = await request.json()
    message = data.get("message", "")
    # Utilisation d'un historique vide pour simplifier
    response_text = generate_response(message, history=[])
    return JSONResponse({"response": response_text})

@app.get("/", response_class=HTMLResponse)
async def root():
    with open("index.html", "r", encoding="utf-8") as f:
        html_content = f.read()
    return HTMLResponse(content=html_content, status_code=200)

if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8000)