import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained(
    'Qwen/Qwen2.5-7B-Instruct',
    dtype=torch.float16,
    device_map='cpu',
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained('Qwen/Qwen2.5-7B-Instruct', trust_remote_code=True)
print('Qwen2.5-7B-Instruct base model downloaded and cached locally for CPU.')