from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Öffentliches, kleines Modell
MODEL_NAME = "Qwen/Qwen2-0.5B"

print("Lade Modell:", MODEL_NAME)
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_NAME,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    device_map="auto"
)
print("Modell bereit.")