tomodrgn.convergence.generate_test_train_pair_volumes#

generate_test_train_pair_volumes(z_train: ndarray, z_test: ndarray, epoch: int, workdir: str, volume_count=100) → None[source]#

Select random particle indices and generate corresponding volumes using train-split images’ latent embeddings and test-split latent embeddings.

Parameters:

  • z_train – numpy array of shape (n_particles, zdim) of latent values deriving from train particle images

  • z_test – numpy array of shape (n_particles) zdim) of latent values deriving from test particle images

  • epoch – epoch being analyzed

  • workdir – str, absolute path to model workdir to store generated volumes

  • volume_count – int, number of volumes to generate for each of train/test

Returns:

None