app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

MAX_HISTORY = 7
MODEL_PATH = 'llongpre/DialoGPT-small-miles'

tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
model = AutoModelForCausalLM.from_pretrained(MODEL_PATH)

# def predict(input, history=[]):
#     # tokenize the new input sentence
#     new_user_input_ids = tokenizer.encode(input + tokenizer.eos_token, return_tensors='pt')
#
#     # append the new user input tokens to the chat history
#     bot_input_ids = torch.cat([torch.LongTensor(history), new_user_input_ids], dim=-1)
#
#     # generate a response
#     history = model.generate(
#                             bot_input_ids,
#                             max_length=1000,
#                             pad_token_id=tokenizer.eos_token_id,
#                             no_repeat_ngram_size=3,
#                             top_p = 0.92,
#                             top_k = 50
#                             ).tolist()
#
#     # convert the tokens to text, and then split the responses into lines
#     response = tokenizer.decode(history[0]).split("<|endoftext|>")
#     response = [(response[i], response[i+1]) for i in range(0, len(response)-1, 2)]  # convert to tuples of list
#
#     return response, history
#
# from transformers.utils import logging

logging.set_verbosity_info()
logger = logging.get_logger("transformers")
logger.info("INFO")

def generate_answer(input, history=[]):
    new_user_input_ids = tokenizer.encode(input + tokenizer.eos_token, return_tensors='pt')
    history = history.append(input)
    logger.info(history)
    if len(history) > MAX_HISTORY:
        history = history[-MAX_HISTORY:]
    bot_input_ids = torch.cat(history, dim=-1)
    chat_history_ids = model.generate(
        bot_input_ids,
        pad_token_id=tokenizer.pad_token_id,
        max_length=1000,
        do_sample=True,
#         top_k=150, # sample from the top k words sorted descending by probability
        top_p=0.7, # choose smallest possible words whose cumulative probability exceeds p
        temperature = 0.95, # 0 greedy, inf is random
        no_repeat_ngram_size=3,
    )
    response = chat_history_ids[:, bot_input_ids.shape[-1]:]
    output = tokenizer.decode(response[0], skip_special_tokens=True)
    history.append(output)

    return output, history


gr.Interface(
    fn=generate_answer,
    title="DialoGPT-large",
    inputs=["text", "state"],
    outputs=["chatbot", "state"],
).launch()