Run train.py file from inside /Repwalk/ directory by providing appropriate arguments(arguments listed in train.py). Word embeddings used is Glove embedding. Extract the file glove.840B.300d.zip to /glove/ folder inside /Repwalk/ folder. Example command given below.
python3 train.py --dataset twitter
Run main.py file from inside /pss/ directory by providing appropriate arguments(arguments listed in train.py). Word embeddings used is Glove embedding. Extract the file glove.840B.300d.zip to root folder of project. Example command given below.
python3 main.py --batch_size 64 --dataset_name Twitter --beta .5
Run train.py file from inside /RepWalk_combined/ directory by providing appropriate arguments(arguments listed in train.py). Word embeddings used is Glove embedding. Extract the file glove.840B.300d.zip to root folder of project. Example commands are given below.
python3 train.py --cpt --dataset laptop --lr .001 --batch_size 32 --phi 10.0 --entropy 2.5 --eps 0.01 --beta 0.01
python3 train.py --cpt --dataset twitter --lr .001 --batch_size 64 --phi 10.0 --entropy 2.0 --eps 0.01 --beta 0.01 --num_epoch 40
python3 train.py --cpt --dataset restaurant --lr .001 --batch_size 8 --phi 10.0 --entropy 2.5 --eps 0.01 --beta 0.01
To run other Variants: --cpt tag is used to run with CPT and --no_itr tag is used to run without Pogressive Attention Supervision.