Document Classification using Transformers

Setup

Prerequisites: Anaconda

1. Set up conda env from the env.yml file.
2. Install english spacy: python -m spacy download en

Conda

To dump the environment: conda env export --no-builds > env.yml

Then, can recreate it somewhere else using: conda env create -f env.yml -n dissertation python=3.9.1

Run

python main.py --config config/dev.json --eval data/processed/fce/json/dev_small.json --train data/processed/fce/json/dev_small.json

Slurm

Change config and datasets in the run_slurm.sh file

Submit job: sbatch run_slurm.sh

Explore BeerAdvocate data

python processed/beer_advocate/explore_dist.py -input_json original/beer_advocate/annotations.json