Perturbations with categorical target data

[bnglomics]

Hi there, Is there a way readily change the *_id_assoc_bayes/ttest.yaml file to account for multiple levels of "drug". For example, I have survey data where the responses are coded as integers from 1 to 5, corresponding to my interventions.. Thank you!!

defaults:

• identify_associations_bayes

batch_size: 10

num_refits: 40

target_dataset: survey_data target_value: 1 <-- how would I change this save_refits: True

model: num_hidden:

• 1000 num_latent: 150

training_loop: lr: 1e-4 num_epochs: 40

[bnglomics]

ok, never mind! I can do one-hot encoding as in the paper :)