model_name = "EdwardYu/llama-2-7b-MedQuAD-merged"

tokenizer = AutoTokenizer.from_pretrained(model_name)

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    load_in_4bit=True,
    torch_dtype=torch.bfloat16,
    device_map="auto",
    quantization_config=BitsAndBytesConfig(
        load_in_4bit=True,
        bnb_4bit_compute_dtype=torch.bfloat16,
        bnb_4bit_use_double_quant=True,
        bnb_4bit_quant_type='nf4'
    ),
)

question = 'What are the side effects or risks of Glucagon?'
inputs = tokenizer(question, return_tensors="pt").to("cuda")
outputs = model.generate(inputs=inputs.input_ids, max_length=1024)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

To run model inference faster, you can load in 16-bits without 4-bit quantization.

python

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
model = PeftModel.from_pretrained(model, adapter)