A fragment-based molecular generation method used GA.

The method could preform multi-objective optimization.

generate_frag_lib.py: Decomposing the real molecules into fragments.

prediction_model.py: Training the odor prediction model.

mol_generator_random.py: Generating new molecules by choosing fragments randomly.

mol_generator_ga.py: Generating new molecules by GA.

Put the prediction model in ./model, and data in ./data.