app.py

import transformers
import numpy as np
import torch
import streamlit as st

from transformers import GPT2Tokenizer, GPT2LMHeadModel

@st.cache(allow_output_mutation=True)
def load_model():
  model_ckpt = "bankholdup/rugpt3_song_writer"
  tokenizer = GPT2Tokenizer.from_pretrained(model_ckpt)
  model = GPT2LMHeadModel.from_pretrained(model_ckpt)
  return tokenizer, model

def set_seed(rng=100000):
    rd = np.random.randint(rng)
    np.random.seed(rd)
    torch.manual_seed(rd)

title = st.title("Загрузка модели")
tokenizer, model = load_model()
title.title("ruGPT3 Song Writer")
context = st.text_input("Введите начало песни", "Как дела? Как дела? Это новый кадиллак")

if st.button("Поехали", help="Может занять какое-то время"):
    generated_sequences = []
    set_seed()
    st.write("Генерируем")
    prompt_text = f"{context}"
    encoded_prompt = tokenizer.encode(prompt_text, add_special_tokens=False, return_tensors="pt")
    output_sequences = model.generate(
            input_ids=encoded_prompt,
            max_length=250 + len(encoded_prompt[0]),
            temperature=1.0,
            top_k=50,
            top_p=0.95,
            repetition_penalty=1.0,
            do_sample=True,
            num_return_sequences=1,
        )
    if len(output_sequences.shape) > 2:
        output_sequences.squeeze_()   
        
    for generated_sequence_idx, generated_sequence in enumerate(output_sequences):
        generated_sequence = generated_sequence.tolist()
        text = tokenizer.decode(generated_sequence, clean_up_tokenization_spaces=True)
    
        #text = text[: text.find("/n") if "/n" else None]

        total_sequence = (
            prompt_text + text[len(tokenizer.decode(encoded_prompt[0], clean_up_tokenization_spaces=True)) :]
        )
        
        reformat_seq = ''.join(['\n'+ x if x.isupper() else x for x in total_sequence ])
    
        #generated_sequences.append(total_sequence)
        st.write(reformat_seq)