license: apache-2.0
language:
- en
tags:
- text-generation
- text2text-generation
pipeline_tag: text2text-generation
widget:
- text: >-
Given the dialog: do you like dance? [SEP] Yes I do. Did you know Bruce
Lee was a cha cha dancer?
example_title: Example1
- text: >-
Given the dialog: i used to scare for darkness [X_SEP] it feels like
hitting to blank wall when i see the darkness [SEP] Oh ya? I don't really
see how [SEP] dont you feel so.. its a wonder [SEP] I do actually hit
blank walls a lot of times but i get by
example_title: Example2
MVP-open-dialog
The MVP-open-dialog model was proposed in MVP: Multi-task Supervised Pre-training for Natural Language Generation by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
The detailed information and instructions can be found https://github.com/RUCAIBox/MVP.
Model Description
MVP-open-dialog is a prompt-based model that MVP is further equipped with prompts pre-trained using labeled open dialogue system datasets. It is a variant (MVP+S) of our main MVP model. It follows a Transformer encoder-decoder architecture with layer-wise prompts.
MVP-open-dialog is specially designed for open dialogue system (conversation) tasks, such as chitchat (PersonaChat, DailyDialog), knowledge grounded conversation (Topical-Chat, Wizard of Wikipedia) and visual dialog (DSTC7-AVSD).
Example
>>> from transformers import MvpTokenizer, MvpForConditionalGeneration
>>> tokenizer = MvpTokenizer.from_pretrained("RUCAIBox/mvp")
>>> model = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mvp-open-dialog")
>>> inputs = tokenizer(
... "Given the dialog: do you like dance? [SEP] Yes I do. Did you know Bruce Lee was a cha cha dancer?",
... return_tensors="pt",
... )
>>> generated_ids = model.generate(**inputs)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
['I did not know that. I did know that Tupac danced ballet in high school.']