Abstract to Title Generation using Sequence2Sequence Models

Here we are using model-t51-base sequenec2sequence model for generation of Paper title from Abstract. The task belong to Text Summarization domain.

Requirements

pip install -r requirements.txt

Usage

For Training Please follow training_and_dataprepration.ipynb

Processing data

python data_processing.py

Training

python training.py

Inference

from InferAbs2Titile import InferAbs2Title
model = InferAbs2Title('m-aliabbas/model-t51-base1')
abstract="""

"""
title = model(abstract)
print(title)

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5.6e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Framework versions

• Transformers 4.27.2
• Pytorch 1.12.1+cu116
• Datasets 2.4.0
• Tokenizers 0.12.1