recipes.cad2.task2.process_dataset package

Submodules

recipes.cad2.task2.process_dataset.generate_train_scenes module

Module to generate new train scenes

recipes.cad2.task2.process_dataset.generate_train_scenes.choose_samples(source: list, number: int) → list

Choose a number of samples from a source list

recipes.cad2.task2.process_dataset.generate_train_scenes.generate_scene_listener(cfg: DictConfig) → None

Generate listeners for the scenes

recipes.cad2.task2.process_dataset.generate_train_scenes.generate_scenes(cfg: DictConfig) → dict

Generate scenes for the training set

recipes.cad2.task2.process_dataset.generate_train_scenes.run(cfg: DictConfig) → None

recipes.cad2.task2.process_dataset.generate_train_scenes.set_song_seed(song: str) → None

Set a seed that is unique for the given song

recipes.cad2.task2.process_dataset.process_zenodo_download module

Module take the data downloaded from Zenodo and rearrange the data for the challenge.

The module expects the data to be downloaded from Zenodo and stored in the following structure:

• zenodo_download_path

  • EnsembleSet.zip

  • CadenzaWoodwind.zip

  • metadata.zip

  • URMP.zip

  • BACH10.zip

  • CadenzaWoodwind.json

  • EnsembleSet.json

recipes.cad2.task2.process_dataset.process_zenodo_download.create_audio_train(cfg)

Create the audio training dataset.

recipes.cad2.task2.process_dataset.process_zenodo_download.create_audio_valid(cfg)

Create the audio validation dataset

recipes.cad2.task2.process_dataset.process_zenodo_download.prepare_cad2_dataset(config: DictConfig) → None

Prepare the dataset for the challenge.

Module contents