models.py

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from config import MODEL_NAME
import spaces

model = None
tokenizer = None

@spaces.GPU
def load_model():
    global model, tokenizer
    if model is None or tokenizer is None:
        tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
        model = AutoModelForCausalLM.from_pretrained(
            MODEL_NAME,
            torch_dtype=torch.float16,
            device_map="auto"
        )
        model.eval()
    return model, tokenizer

@spaces.GPU
def generate_response(message, history, max_length=512, temperature=0.7):
    model, tokenizer = load_model()
    
    # Prepare input
    if history:
        input_text = history + f"\nUser: {message}\nAssistant:"
    else:
        input_text = f"User: {message}\nAssistant:"
    
    inputs = tokenizer(input_text, return_tensors="pt").to(model.device)
    
    with torch.no_grad():
        outputs = model.generate(
            **inputs,
            max_length=max_length,
            temperature=temperature,
            do_sample=True,
            pad_token_id=tokenizer.eos_token_id
        )
    
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    # Extract only the assistant's response
    if "Assistant:" in response:
        response = response.split("Assistant:")[-1].strip()
    return response