from gptqmodel import GPTQModel
from transformers import AutoTokenizer

model     = GPTQModel.load("atrisaxena/llama3.1-8B-gptq-4bit", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("atrisaxena/llama3.1-8B-gptq-4bit")

messages = [{"role": "user", "content": "What is quantization?"}]
prompt   = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs   = tokenizer(prompt, return_tensors="pt").to(model.device)

output = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(output[0], skip_special_tokens=True))