from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import torch
from transformers import GenerationConfig

use_cuda = torch.cuda.is_available()
device = torch.device("cuda" if use_cuda else "cpu")
generation_config = GenerationConfig.from_pretrained("Den4ikAI/FRED-T5-LARGE_text_qa")
tokenizer = AutoTokenizer.from_pretrained("Den4ikAI/FRED-T5-LARGE_text_qa")
model = AutoModelForSeq2SeqLM.from_pretrained("Den4ikAI/FRED-T5-LARGE_text_qa").to(device)
model.eval()

def generate(prompt):
  data = tokenizer(f"{prompt}", return_tensors="pt").to(model.device)
  output_ids = model.generate(
      **data,
      generation_config=generation_config
  )[0]
  print(tokenizer.decode(data["input_ids"][0].tolist()))
  out = tokenizer.decode(output_ids.tolist())
  return out

while 1:
  prompt = '''<SC6>Текст: {}\nВопрос: {}\nОтвет: <extra_id_0>
  '''.format(input('Текст: '), input('Вопрос: '))
  print(generate(prompt))