Variational Representation Learning by using proximity loss functions
The data and saved models can be downloaded from: resources
Replicate the conda environment in the environment file using: conda env create -f environment.yml
Train the model using the following command:
allennlp train -s models/info_vae_snli_cosine_reg --include-package src \
config/info_Vae_snli_cosine_reg.jsonnet
Can perform transfer using the following command:
python prox_vae.py transfer models/info_vae_snli_cosine_reg --templates_file \
data/interim/snli_1.0/templates.tsv