This repository consists on a Pytorch implementation of TableNet.
To training or predict, you should first install the requirements by running the following code:
pip install -r requirements.txt
To train is only needed the train.py
file which can be configured as wanted.
marmot.py
and tablenet.py
are inheritance of Pytorch Lighting modules: LightningDataModule
and LightningModule
, respectively.
To predict, it can be used the pre-trained weights already available and should be downloaded on the following link: TableNet Weights
python predict.py --model_weights='<weights path>' --image_path='<image path>'
or simply:
python predict.py
To predict with the default image.