app.py

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

tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-large")
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-large")

def chat_with_history(message, chat_history=None):
    # Initialize chat history if not provided
    if chat_history is None:
        chat_history = []

    # Encode the new user input, add the eos_token, and return a tensor in PyTorch
    new_user_input_ids = tokenizer.encode(message + tokenizer.eos_token, return_tensors='pt')

    # Append the new user input tokens to the chat history
    bot_input_ids = torch.cat([tokenizer.encode(pair[0] + tokenizer.eos_token, return_tensors='pt') for pair in chat_history] + [new_user_input_ids], dim=-1)

    # Generate a response while limiting the total chat history to 1000 tokens
    chat_history_ids = model.generate(bot_input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id)

    # Decode the last output tokens from bot
    response = tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)

    # Update the chat history with the new user message and bot response
    chat_history.append([message, response])

    return response

demo = gr.ChatInterface(
        fn=chat_with_history,
        examples=["hey how are you ?", "hola", "Yo!"],
        title="Multi Chat Bot"
    )

demo.launch()