Datasets for manuscript - Dirichlet diffusion score model for biological sequence generation.

This repository holds the trained Dirichlet Diffusion Score models for various datasets.

best_models.tar.gz

It also contains all input data required to train your own models with scripts provided via github repository.

data.tar.gz

This archive contains the following folders: 

satnet_sudoku contains dataset with sudoku examples which we used for evaluation of sudoku model.

promoter_design contains dataset used for training promoter design model as well as Sei model weights. Please, read provided readme file before using it for training scripts.