Parameters
--model vae (default), pairwise, cost
--retrain 'path/to/data', (default) ' '
--name 'name-of-themodel', (default) ' '
--lr '[1e-6, 1e-3]', (default) '[1e-6, 0.1]'
--epochs 10, (default) 100
--trials 15, (default) 25
Example use model.py --model <model_name> --retrain --model-path <model_path>
python main.py --model pairwise --retrain src/Data/pairwise-encodings.txt --model-path src/Data/vae.onnx --name model-nameThanks to Ryan Marcus for the implementation of "Tree Convolution" that was used in the training script.