from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig
from utils import get_gpu_memory_info

device = "cuda" # the device to load the model onto

model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen2-7B-Instruct",
    torch_dtype="auto",
    device_map="auto"
)

tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-7B-Instruct")

def respond(
    message,
):
    messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": message}
]
    get_gpu_memory_info()
    text = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )
    model_inputs = tokenizer([text], return_tensors="pt").to(device)

    generated_ids = model.generate(
        model_inputs.input_ids,
        max_new_tokens=512
    )
    generated_ids = [
        output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
    ]

    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
    get_gpu_memory_info()
    return response