app.py

from fastapi import FastAPI
from pydantic import BaseModel
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

app = FastAPI()

# Chargement du modèle
model = AutoModelForCausalLM.from_pretrained("deepseek-ai/deepseek-llm-7b-chat", device_map="auto", torch_dtype=torch.float16)
tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/deepseek-llm-7b-chat")

class Prompt(BaseModel):
    prompt: str

@app.post("/predict")
async def predict(prompt: Prompt):
    inputs = tokenizer(prompt.prompt, return_tensors="pt").to(model.device)
    outputs = model.generate(**inputs, max_new_tokens=256, do_sample=True)
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return {"response": response}