TextToTable

A tool for transforming text to table using a model trained with shirley-wu/text_to_table.

It is a project for the assignment of Fundations of Data Science (90213102@NJU).

How to use

First clone the project and open the project directory.

git clone https://github.com/wsq4/TextToTable.git
cd TextToTable

For preprocessing, we use fairseq for BPE and binarization. You need to first download a BART (bart.base) model here, our preprocessed dataset here and our pretrained model here. Then extract and put them in the project directory.

Then use pipenv to create an virtual Python 3.7 environment and install dependencies.

pipenv --python 3.7
pipenv install

After that, you can enter the virtual environment to run the program.

pipenv shell
bash bash ./text-to-table.sh Input.txt ./output/