app.py

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

app = FastAPI()

model_name = "Xennus/niko-mistral-cyberbot"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)
model.eval()

class Request(BaseModel):
    prompt: str

@app.post("/generate")
async def generate(request: Request):
    inputs = tokenizer(request.prompt, return_tensors="pt")
    outputs = model.generate(**inputs, max_length=200)
    text = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return {"response": text}