torchdrivesim.rendering.pytorch3d

PyTorch3D-based renderers, used by default in TorchDriveSim.

Attributes

is_available

Exceptions

Pytorch3DNotFound

Import can't find module, or can't find name in module.

Classes

RenderingBlend	Blending choices from pytorch3d. May be ignored by other renderer types.
Pytorch3DRendererConfig	Configuration of pytorch3d-based renderer.
Shader2D	Shader that ignores lighting, based on https://github.com/facebookresearch/pytorch3d/issues/84
Pytorch3DRenderer	Renderer based on pytorch3d, using an orthographic projection and a trivial shader.

Functions

construct_pytorch3d_cameras(...)

Module Contents

torchdrivesim.rendering.pytorch3d.is_available = True

exception torchdrivesim.rendering.pytorch3d.Pytorch3DNotFound

Bases: ImportError

Import can’t find module, or can’t find name in module.

class torchdrivesim.rendering.pytorch3d.RenderingBlend

Bases: enum.Enum

Blending choices from pytorch3d. May be ignored by other renderer types. https://pytorch3d.readthedocs.io/en/latest/modules/renderer/blending.html

hard = 'hard'

soft = 'soft'

sigmoid = 'sigmoid'

class torchdrivesim.rendering.pytorch3d.Pytorch3DRendererConfig

Bases: torchdrivesim.rendering.base.RendererConfig

Configuration of pytorch3d-based renderer.

backend: str = 'pytorch3d'

differentiable_rendering: RenderingBlend

class torchdrivesim.rendering.pytorch3d.Shader2D(device='cpu', background_color: Tuple[float, float, float] = (0, 0, 0), blend=RenderingBlend.soft)

Bases: torch.nn.Module

Shader that ignores lighting, based on https://github.com/facebookresearch/pytorch3d/issues/84

forward(fragments, meshes, **kwargs) → torch.Tensor

class torchdrivesim.rendering.pytorch3d.Pytorch3DRenderer(cfg: Pytorch3DRendererConfig, *args, **kwargs)

Bases: torchdrivesim.rendering.base.BirdviewRenderer

Renderer based on pytorch3d, using an orthographic projection and a trivial shader. Works on both GPU and CPU, but CPU is very slow.

render_mesh(mesh: torchdrivesim.mesh.BirdviewMesh, res: torchdrivesim.utils.Resolution, cameras: torchdrivesim.rendering.base.Cameras) → torch.Tensor

Renders a given mesh, producing BxHxWxC tensor image of float RGB values in [0,255] range.

classmethod make_renderer(res, blend, background_color)

torchdrivesim.rendering.pytorch3d.construct_pytorch3d_cameras(cameras: torchdrivesim.rendering.base.Cameras) → pytorch3d.renderer.FoVOrthographicCameras