---
thumbnail: https://huggingface.co/front/thumbnails/dialogpt.png
language:
- en
license: cc-by-4.0
tags:
- conversational
- transformers
datasets:
- multi_woz_v22 and PIE-English idioms corpus
metrics:
- perplexity
widget:
- text: "Does that mean Jane is off the hook?"
---

## DialoGPT_MWOZ_Idioms

This is a fine-tuned model of DialoGPT (medium)-MultiWOZ on the PIE-English idioms corpus. It is intended to be used as an idiom-aware conversational system.
The dataset it's trained on is limited in scope, as it covers only 10 classes of idioms ( metaphor, simile, euphemism, parallelism, personification, oxymoron, paradox, hyperbole, irony and literal).

The perplexity achieved on the test and validation sets are 160.9 and 160.46, respectively. More information about the original pre-trained model can be found [here](https://huggingface.co/tosin/dialogpt_mwoz)

* Generation example from an interactive environment:

|Role | Response |
|---------|------------|
|User | remind her to take it off the hook. |
|Bot | the bookie took it on the nod |
|User | A lengthy break is in order ; so make sure you keep your nose clean |
|Bot | they are under a cloud because of fears about US President's wife Hillary Clinton |
|User | Some families breed like rabbits and send us all their male heirs |
|Bot | he was out for the count |

Please find the information about preprocessing, training and full details of the DialoGPT in the [original DialoGPT repository](https://github.com/microsoft/DialoGPT)

### How to use
Now we are ready to try out how the model works as a chatting partner!

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
tokenizer = AutoTokenizer.from_pretrained("tosin/dialogpt_mwoz_idioms")
model = AutoModelForCausalLM.from_pretrained("tosin/dialogpt_mwoz_idioms")
# 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_MWOZ_Bot: {}".format(tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)))