import gradio as gr
from huggingface_hub import hf_hub_download
import torch
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
tokenizer = AutoTokenizer.from_pretrained("VietAI/vit5-base")
model = AutoModelForSeq2SeqLM.from_pretrained("VietAI/vit5-base")
model_file = hf_hub_download(repo_id="tuongvxx1/medBot", filename="medicalBot_ver2.pth")
model.load_state_dict(torch.load(model_file, map_location=torch.device('cpu')))
model.to(device)

def generate_answer(question, model, tokenizer, device):
    model.eval()
    input_text = "hỏi: " + question
    inputs = tokenizer(input_text, return_tensors="pt", max_length=512, truncation=True, padding="max_length")
    input_ids = inputs.input_ids.to(device)
    attention_mask = inputs.attention_mask.to(device)

    with torch.no_grad():
        outputs = model.generate(input_ids=input_ids, attention_mask=attention_mask, max_length=128, num_beams=4, early_stopping=True)
    answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return answer

def run(ques):
    return generate_answer(ques, model, tokenizer, device)

demo = gr.Interface(fn=run, inputs=gr.Textbox(label="Nhập câu hỏi"), outputs=gr.Textbox(label="Câu trả lời"))
demo.launch()