Move neural nets factory out of utils (#994)

* Export classes and function

* Reworked factory imports

* Debugging ruff

* Debugging ruff

* Added links to the new location with deprecation warnings

* Forgot build_zuko_maf

* Added tests for the deprecated functions

* Renaming

* Doc strings now only contain the deprecation warning

* Updated the deprecation message

* Updated the deprecation message

* Ruff once more

docs/docs/reference.md

@@ -116,17 +116,17 @@
 
 ## Models
 
-::: sbi.utils.get_nn_models.posterior_nn
+::: sbi.neural_nets.factory.posterior_nn
     rendering:
       show_root_heading: true
       show_object_full_path: true
 
-::: sbi.utils.get_nn_models.likelihood_nn
+::: sbi.neural_nets.factory.likelihood_nn
     rendering:
       show_root_heading: true
       show_object_full_path: true
 
-::: sbi.utils.get_nn_models.classifier_nn
+::: sbi.neural_nets.factory.classifier_nn
     rendering:
       show_root_heading: true
       show_object_full_path: true

sbi/inference/snle/snle_base.py

@@ -12,11 +12,10 @@
 from torch.nn.utils.clip_grad import clip_grad_norm_
 from torch.utils.tensorboard.writer import SummaryWriter
 
-from sbi import utils as utils
 from sbi.inference import NeuralInference
 from sbi.inference.posteriors import MCMCPosterior, RejectionPosterior, VIPosterior
 from sbi.inference.potentials import likelihood_estimator_based_potential
-from sbi.neural_nets.density_estimators import DensityEstimator
+from sbi.neural_nets import DensityEstimator, likelihood_nn
 from sbi.utils import check_estimator_arg, check_prior, x_shape_from_simulation
 
 
@@ -64,7 +63,7 @@ def __init__(
         # potentially for z-scoring.
         check_estimator_arg(density_estimator)
         if isinstance(density_estimator, str):
-            self._build_neural_net = utils.likelihood_nn(model=density_estimator)
+            self._build_neural_net = likelihood_nn(model=density_estimator)
         else:
             self._build_neural_net = density_estimator
 

sbi/inference/snpe/snpe_base.py

@@ -22,7 +22,7 @@
 )
 from sbi.inference.posteriors.base_posterior import NeuralPosterior
 from sbi.inference.potentials import posterior_estimator_based_potential
-from sbi.neural_nets.density_estimators import DensityEstimator
+from sbi.neural_nets import DensityEstimator, posterior_nn
 from sbi.utils import (
     RestrictedPrior,
     check_estimator_arg,
@@ -77,7 +77,7 @@ def __init__(
         # potentially for z-scoring.
         check_estimator_arg(density_estimator)
         if isinstance(density_estimator, str):
-            self._build_neural_net = utils.posterior_nn(model=density_estimator)
+            self._build_neural_net = posterior_nn(model=density_estimator)
         else:
             self._build_neural_net = density_estimator
 

sbi/inference/snpe/snpe_c.py

@@ -51,9 +51,9 @@ def __init__(
         In this codebase, we will automatically switch to the non-atomic loss if the
         following criteria are fulfilled:<br/>
         - proposal is a `DirectPosterior` with density_estimator `mdn`, as built
-            with `utils.sbi.posterior_nn()`.<br/>
+            with `sbi.neural_nets.posterior_nn()`.<br/>
         - the density estimator is a `mdn`, as built with
-            `utils.sbi.posterior_nn()`.<br/>
+            `sbi.neural_nets.posterior_nn()`.<br/>
         - `isinstance(prior, MultivariateNormal)` (from `torch.distributions`) or
             `isinstance(prior, sbi.utils.BoxUniform)`
 

sbi/inference/snre/snre_base.py

@@ -12,6 +12,7 @@
 from sbi.inference.base import NeuralInference
 from sbi.inference.posteriors import MCMCPosterior, RejectionPosterior, VIPosterior
 from sbi.inference.potentials import ratio_estimator_based_potential
+from sbi.neural_nets import classifier_nn
 from sbi.utils import (
     check_estimator_arg,
     check_prior,
@@ -74,7 +75,7 @@ def __init__(
         # potentially for z-scoring.
         check_estimator_arg(classifier)
         if isinstance(classifier, str):
-            self._build_neural_net = utils.classifier_nn(model=classifier)
+            self._build_neural_net = classifier_nn(model=classifier)
         else:
             self._build_neural_net = classifier
 

sbi/neural_nets/__init__.py

@@ -0,0 +1,23 @@
+from sbi.neural_nets.classifier import (
+    StandardizeInputs,
+    build_input_layer,
+    build_linear_classifier,
+    build_mlp_classifier,
+    build_resnet_classifier,
+)
+from sbi.neural_nets.density_estimators import DensityEstimator, NFlowsFlow
+from sbi.neural_nets.embedding_nets import (
+    CNNEmbedding,
+    FCEmbedding,
+    PermutationInvariantEmbedding,
+)
+from sbi.neural_nets.factory import classifier_nn, likelihood_nn, posterior_nn
+from sbi.neural_nets.flow import (
+    build_made,
+    build_maf,
+    build_maf_rqs,
+    build_nsf,
+    build_zuko_maf,
+)
+from sbi.neural_nets.mdn import build_mdn
+from sbi.neural_nets.mnle import MixedDensityEstimator, build_mnle