dargilboa / mdma Goto Github PK

I am trying to do a conditional sampling using the toy example(gaussian) , however i got this error

n_sample = 200
condX = np.array([[0,1]])
model.sample(n_sample,inds = [2],cond_inds = [0,1],cond_X = t.tensor(condX))

hello,

congrats on the interesting work/package! i am interested in trying your approach to:

1. fit a density model, say p(x_1, x_2, x_3, x_4)

and

2. sample from univariate conditionals like p(x_2 | x_1, x_3, x_4)

however, i'm at bit confused about the API. how can i go about doing this?

thanks!