- Extract the files from
Pokemon/original_data.zip
- Run the preprocessing script
../preprocess.py
on the dataset. - If you need images of Pokemon fusions, follow steps 4-7. Otherwise stop here.
- Download the fusion dataset from https://github.com/Aegide/FusionSprites
- Run
../scripts/data/fix_infinite_fusion_filenames.py
on theJapeal
folder within that repository. - Run
../scripts/data/split_fusion_data.py
on the folder from above and the extracted files from step 1. - Run
../scripts/data/split_training_fusion_data.py
on the two training directories resulting from the above.
- Download and extract the dataset from https://veekun.com/static/pokedex/downloads/pokemon-sugimori.tar.gz
- Remove any empty images in the extracted directory.
- Run the preprocessing script
../preprocess.py
on the dataset.
- Clone the repository from https://github.com/YingzhenLi/Sprites.
- Run the
random_character.py
in this codebase. - Pass the
frames
directory generated from the above to../scripts/data/process_sprite_dataset_files.py
- Run the preprocessing script
../preprocess.py
on the dataset.
- Extract and download the files from https://github.com/AgaMiko/pixel_character_generator.
- Run
rename_tiny_hero_files.py
on the dataset. - Run the preprocessing script
../preprocess.py
on the dataset.
https://www.kaggle.com/mariotormo/complete-pokemon-dataset-updated-090420
File: pokedex_(Update_04.21).csv
This dataset is used in the following scripts:
fusion-dance\scripts\compute_vae_embeddings.py
fusion-dance\scripts\compute_vqvae_embeddings.py
fusion-dance\scripts\data\create_classification_labels.py
fusion-dance\scripts\data\create_pokemon_metadata.py
https://github.com/PokeAPI/pokeapi
Files: pokemon_colors.csv
, pokemon_shapes.csv
, pokemon_species.csv
These datasets are used in the following script:
fusion-dance\scripts\data\create_pokemon_metadata.py