app.py

from transformers import AutoModelForCausalLM, AutoTokenizer  
import gradio as gr  
import torch  
  
title = "EZChat"  
description = "A State-of-the-Art Large-scale Pretrained Response generation model (DialoGPT-medium)"  
examples = [["How are you?"]]  
  
# Set the padding token to be used and initialize the model  
tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-medium")
tokenizer.padding_side = 'left'  
tokenizer.add_special_tokens({'pad_token': '[EOS]'})  
tokenizer.pad_token = tokenizer.eos_token  

# Model
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-medium")  
  
#predict  
def predict(input, history=[]):  
    # tokenize the new input sentence  
    new_user_input_ids = tokenizer.encode(  
        input + tokenizer.eos_token, padding=True, truncation=True, return_tensors="pt"  
    )  
  
    # append the new user input tokens to the chat history  
    bot_input_ids = torch.cat([torch.tensor(history), new_user_input_ids], dim=-1) if history else new_user_input_ids  
  
    # generate a response  
    chat_history_ids = model.generate(  
        bot_input_ids, max_length=4000, pad_token_id=tokenizer.eos_token_id  
    )  
  
    # convert the tokens to text, and then split the responses into lines  
    response = tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)  
      
    return response, chat_history_ids.tolist()[0]  
  
  
iface = gr.Interface(  
    fn=predict,  
    title=title,  
    description=description,  
    examples=examples,  
    inputs=["text", gr.inputs.Slider(0, 4000, default=2000, label='Chat History')],  
    outputs=["text", "text"],  
    theme="ParityError/Anime",  
)  
  
iface.launch()