How to train without integrating¶

ZüNIS allows to train a model without performing the integration. For this, we can use the trainer API:

import torch
from zunis.models.flows.sampling import UniformSampler
from zunis.training.weighted_dataset.stateful_trainer import StatefulTrainer

device = torch.device("cuda")

d = 2

def f(x):
    return x[:,0]**2 + x[:,1]**2

trainer = StatefulTrainer(d=d, device=device)
x, px, fx=trainer.generate_target_batch_from_posterior(10, f, UniformSampler(d=d, device=device))
trainer.train_on_batch(x,px,fx)

In a first step, we initialize a trainer object. We generate a batch of 10 uniformly sampled points in the 2D hypercube x with the probability distribution px and the function value fx. Then, one training step is performed on this batch.