SMC-decoding

requirements

pip install -r requirements.txt
Install the datasets library from source:
cd ..
git clone https://github.com/huggingface/datasets.git
cd datasets
pip install -e .

if not working:

install cython and scm_tools: *pip install cython
- pip install setuptools_scm
upgrade pip

Before running scripts: export PYTHONPATH=src:${PYTHONPATH}

Loading and saving cached dataset and models

python src/save_datasets_models.py

Training the attribute model

with GPT2 Tokenizer

python src/attribute_models/train_attribute_model.py -out_path "output/sst_attribute_model" -model "lstm" -label 1 -emb_size 32 -hidden_size 64 -p_drop 0.0 -ep 50 -bs 32 -num_workers 4
python src/attribute_models/train_attribute_model.py -out_path "output/sst_attribute_model" -model "gpt2" -label 1 -p_drop 0.0 -ep 50 -bs 32 -num_workers 4

with SST Tokenizer

python src/attribute_models/train_attribute_model.py -out_path "output/sst_attribute_model" -model "lstm" -tokenizer "sst" -label 1 -emb_size 32 -hidden_size 64 -p_drop 0.0 -ep 5 -bs 32 -num_workers 4