Paper Title Generator

Generates titles for computer science papers given an abstract.

The model is a BERT2BERT Encoder-Decoder using the official bert-base-uncased checkpoint as initialization for the encoder and decoder. It was fine-tuned on 318,500 computer science papers posted on arXiv.org between 2007 and 2022 and achieved a 26.3% Rouge2 F1-Score on held-out validation data.

Live Demo: https://paper-titles.ey.r.appspot.com/

Downloads last month
190
Safetensors
Model size
247M params
Tensor type
I64
ยท
F32
ยท
Inference Providers NEW
This model is not currently available via any of the supported third-party Inference Providers, and the model is not deployed on the HF Inference API.

Dataset used to train Callidior/bert2bert-base-arxiv-titlegen

Spaces using Callidior/bert2bert-base-arxiv-titlegen 4