transformers-into-vaes

Code for Finetuning Pretrained Transformers into Variational Autoencoders (our submission to NLP Insights Workshop 2021).

Gathering data used in the paper:

  1. Download all data (penn, snli, yahoo, yelp) from this repository.

  2. Change data path in base_models.py accordingly.

Running experiments:

  1. Install dependencies.
pip install -r requirements.txt
  1. Run phase 1 (encoder only training):
./run_encoder_training snli
  1. Run phase 2 (full training):
./run_training snli <path_to_checkpoint_from_phase_1>

Calculating metrics:

python evaluate_all.py -d snli -bs 256 -c <path_to_config_file> -ckpt <path_to_checkpoint_file>

GitHub

https://github.com/seongminp/transformers-into-vaes