clean renderer args

Summary: continued - don't duplicate inputs

Reviewed By: kjchalup

Differential Revision: D38248829

fbshipit-source-id: 2d56418ecbec9cc597c3cf0c122199e274661516

projects/implicitron_trainer/configs/repro_multiseq_idr_ad.yaml

@@ -43,7 +43,6 @@ model_factory_ImplicitronModelFactory_args:
         line_step_iters: 3
         n_secant_steps: 8
         n_steps: 100
-        object_bounding_sphere: 8.0
         sdf_threshold: 5.0e-05
       ray_normal_coloring_network_args:
         d_in: 9

projects/implicitron_trainer/configs/repro_singleseq_idr.yaml

@@ -36,7 +36,6 @@ model_factory_ImplicitronModelFactory_args:
         line_step_iters: 3
         n_secant_steps: 8
         n_steps: 100
-        object_bounding_sphere: 8.0
         sdf_threshold: 5.0e-05
       ray_normal_coloring_network_args:
         d_in: 9

projects/implicitron_trainer/tests/experiment.yaml

@@ -241,15 +241,6 @@ model_factory_ImplicitronModelFactory_args:
         density_relu: true
         blend_output: false
     renderer_SignedDistanceFunctionRenderer_args:
-      render_features_dimensions: 3
-      ray_tracer_args:
-        object_bounding_sphere: 1.0
-        sdf_threshold: 5.0e-05
-        line_search_step: 0.5
-        line_step_iters: 1
-        sphere_tracing_iters: 10
-        n_steps: 100
-        n_secant_steps: 8
       ray_normal_coloring_network_args:
         feature_vector_size: 3
         mode: idr
@@ -266,6 +257,13 @@ model_factory_ImplicitronModelFactory_args:
       bg_color:
       - 0.0
       soft_mask_alpha: 50.0
+      ray_tracer_args:
+        sdf_threshold: 5.0e-05
+        line_search_step: 0.5
+        line_step_iters: 1
+        sphere_tracing_iters: 10
+        n_steps: 100
+        n_secant_steps: 8
     image_feature_extractor_ResNetFeatureExtractor_args:
       name: resnet34
       pretrained: true

pytorch3d/implicitron/models/generic_model.py

@@ -641,35 +641,32 @@ class GenericModel(ImplicitronModelBase):  # pyre-ignore: 13
             **raysampler_args, **extra_args
         )
 
+    @classmethod
+    def renderer_tweak_args(cls, type, args: DictConfig) -> None:
+        """
+        We don't expose certain fields of the renderer because we want to set
+        them based on other inputs.
+        """
+        args.pop("render_features_dimensions", None)
+        args.pop("object_bounding_sphere", None)
+
     def create_renderer(self):
-        raysampler_args = getattr(
-            self, "raysampler_" + self.raysampler_class_type + "_args"
-        )
-        self.renderer_MultiPassEmissionAbsorptionRenderer_args[
-            "stratified_sampling_coarse_training"
-        ] = raysampler_args["stratified_point_sampling_training"]
-        self.renderer_MultiPassEmissionAbsorptionRenderer_args[
-            "stratified_sampling_coarse_evaluation"
-        ] = raysampler_args["stratified_point_sampling_evaluation"]
-        self.renderer_SignedDistanceFunctionRenderer_args[
-            "render_features_dimensions"
-        ] = self.render_features_dimensions
+        extra_args = {}
 
         if self.renderer_class_type == "SignedDistanceFunctionRenderer":
-            if "scene_extent" not in raysampler_args:
+            extra_args["render_features_dimensions"] = self.render_features_dimensions
+            if not hasattr(self.raysampler, "scene_extent"):
                 raise ValueError(
                     "SignedDistanceFunctionRenderer requires"
                     + " a raysampler that defines the 'scene_extent' field"
                     + " (this field is supported by, e.g., the adaptive raysampler - "
                     + " self.raysampler_class_type='AdaptiveRaySampler')."
                 )
-            self.renderer_SignedDistanceFunctionRenderer_args.ray_tracer_args[
-                "object_bounding_sphere"
-            ] = self.raysampler_AdaptiveRaySampler_args["scene_extent"]
+            extra_args["object_bounding_sphere"] = self.raysampler.scene_extent
 
         renderer_args = getattr(self, "renderer_" + self.renderer_class_type + "_args")
         self.renderer = registry.get(BaseRenderer, self.renderer_class_type)(
-            **renderer_args
+            **renderer_args, **extra_args
         )
 
     def create_implicit_function(self) -> None:

pytorch3d/implicitron/models/renderer/multipass_ea.py

@@ -53,10 +53,12 @@ class MultiPassEmissionAbsorptionRenderer(  # pyre-ignore: 13
             fine rendering pass during training.
         n_pts_per_ray_fine_evaluation: The number of points sampled per ray for the
             fine rendering pass during evaluation.
-        stratified_sampling_coarse_training: Enable/disable stratified sampling during
-            training.
-        stratified_sampling_coarse_evaluation: Enable/disable stratified sampling during
-            evaluation.
+        stratified_sampling_coarse_training: Enable/disable stratified sampling in the
+            refiner during training. Only matters if there are multiple implicit
+            functions (i.e. in GenericModel if num_passes>1).
+        stratified_sampling_coarse_evaluation: Enable/disable stratified sampling in
+            the refiner during evaluation. Only matters if there are multiple implicit
+            functions (i.e. in GenericModel if num_passes>1).
         append_coarse_samples_to_fine: Add the fine ray points to the coarse points
             after sampling.
         density_noise_std_train: Standard deviation of the noise added to the

pytorch3d/implicitron/models/renderer/sdf_renderer.py

@@ -8,7 +8,11 @@ from typing import List, Optional, Tuple
 
 import torch
 from omegaconf import DictConfig
-from pytorch3d.implicitron.tools.config import get_default_args_field, registry
+from pytorch3d.implicitron.tools.config import (
+    get_default_args_field,
+    registry,
+    run_auto_creation,
+)
 from pytorch3d.implicitron.tools.utils import evaluating
 from pytorch3d.renderer import RayBundle
 
@@ -18,9 +22,10 @@ from .rgb_net import RayNormalColoringNetwork
 
 
 @registry.register
-class SignedDistanceFunctionRenderer(BaseRenderer, torch.nn.Module):
+class SignedDistanceFunctionRenderer(BaseRenderer, torch.nn.Module):  # pyre-ignore[13]
     render_features_dimensions: int = 3
-    ray_tracer_args: DictConfig = get_default_args_field(RayTracing)
+    object_bounding_sphere: float = 1.0
+    ray_tracer: RayTracing
     ray_normal_coloring_network_args: DictConfig = get_default_args_field(
         RayNormalColoringNetwork
     )
@@ -37,8 +42,7 @@ class SignedDistanceFunctionRenderer(BaseRenderer, torch.nn.Module):
                 f"Background color should have {render_features_dimensions} entries."
             )
 
-        self.ray_tracer = RayTracing(**self.ray_tracer_args)
-        self.object_bounding_sphere = self.ray_tracer_args.get("object_bounding_sphere")
+        run_auto_creation(self)
 
         self.ray_normal_coloring_network_args[
             "feature_vector_size"
@@ -49,6 +53,17 @@ class SignedDistanceFunctionRenderer(BaseRenderer, torch.nn.Module):
 
         self.register_buffer("_bg_color", torch.tensor(self.bg_color), persistent=False)
 
+    @classmethod
+    def ray_tracer_tweak_args(cls, type, args: DictConfig) -> None:
+        del args["object_bounding_sphere"]
+
+    def create_ray_tracer(self) -> None:
+        self.ray_tracer = RayTracing(
+            # pyre-ignore[32]
+            **self.ray_tracer_args,
+            object_bounding_sphere=self.object_bounding_sphere,
+        )
+
     def requires_object_mask(self) -> bool:
         return True
 
@@ -97,7 +112,6 @@ class SignedDistanceFunctionRenderer(BaseRenderer, torch.nn.Module):
         object_mask = object_mask.reshape(batch_size, -1)
 
         with torch.no_grad(), evaluating(implicit_function):
-            # pyre-fixme[29]: `Union[torch.Tensor, torch.nn.Module]` is not a function.
             points, network_object_mask, dists = self.ray_tracer(
                 sdf=lambda x: implicit_function(x)[
                     :, 0
@@ -128,7 +142,6 @@ class SignedDistanceFunctionRenderer(BaseRenderer, torch.nn.Module):
             N = surface_points.shape[0]
 
             # Sample points for the eikonal loss
-            # pyre-fixme[9]
             eik_bounding_box: float = self.object_bounding_sphere
             n_eik_points = batch_size * num_pixels // 2
             eikonal_points = torch.empty(