Dialogpt Small

🚀 A State-of-the-Art Large-scale Pretrained Response generation model (DialoGPT)

DialoGPT is a state-of-the-art large-scale pretrained dialogue response generation model designed for multiturn conversations. It has shown comparable response quality to humans in single-turn conversation Turing tests, as indicated by human evaluation results.

✨ Features

• DialoGPT is a SOTA large-scale pretrained dialogue response generation model for multiturn conversations.
• The human evaluation results show that its single - turn responses are comparable to human responses in quality.
• It is trained on 147M multi - turn dialogues from Reddit discussion threads.

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

Basic Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch


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

# Let's chat for 5 lines
for step in range(5):
    # encode the new user input, add the eos_token and return a tensor in Pytorch
    new_user_input_ids = tokenizer.encode(input(">> User:") + tokenizer.eos_token, return_tensors='pt')

    # append the new user input tokens to the chat history
    bot_input_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1) if step > 0 else new_user_input_ids

    # generated 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)

    # pretty print last ouput tokens from bot
    print("DialoGPT: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))

📚 Documentation

• Multi - turn generation examples from an interactive environment:

• Please find the information about preprocessing, training and full details of the DialoGPT in the original DialoGPT repository
• ArXiv paper: https://arxiv.org/abs/1911.00536
• Open LLM Leaderboard Evaluation Results
  • Detailed results can be found here

📄 License

The model is licensed under the MIT license.