torchdrivesim.rendering.nvdiffrast#

Nvdiffrast-based renderers, equivalent to those based on PyTorch3D but sometimes faster. This module imports correctly if nvdiffrast is missing, but the renderer will raise the NvdiffrastNotFound exception.

Module Contents#

Classes#

NvdiffrastRendererConfig

Configuration of nvdiffrast-based renderer.

NvdiffrastRenderer

Similar to PyTorch3DRenderer, and producing indistinguishable images, but sometimes faster.

Functions#

get_glctx_session(device[, opengl])

Attributes#

is_available

logger

glctx_sessions

torchdrivesim.rendering.nvdiffrast.is_available = True[source]#
torchdrivesim.rendering.nvdiffrast.logger[source]#
torchdrivesim.rendering.nvdiffrast.glctx_sessions[source]#
exception torchdrivesim.rendering.nvdiffrast.NvdiffrastNotFound[source]#

Bases: ImportError

Nvdiffrast is not installed.

torchdrivesim.rendering.nvdiffrast.get_glctx_session(device, opengl=True)[source]#
class torchdrivesim.rendering.nvdiffrast.NvdiffrastRendererConfig[source]#

Bases: torchdrivesim.rendering.base.RendererConfig

Configuration of nvdiffrast-based renderer.

backend: str = 'nvdiffrast'[source]#
antialias: bool = False[source]#
opengl: bool = True[source]#
max_minibatch_size: int | None[source]#
class torchdrivesim.rendering.nvdiffrast.NvdiffrastRenderer(cfg: NvdiffrastRendererConfig, *args, **kwargs)[source]#

Bases: torchdrivesim.rendering.base.BirdviewRenderer

Similar to PyTorch3DRenderer, and producing indistinguishable images, but sometimes faster. Note that nvdiffrast requires separate installation and is subject to its own license terms.

render_mesh(mesh: torchdrivesim.mesh.BirdviewMesh, res: torchdrivesim.utils.Resolution, cameras: pytorch3d.renderer.FoVOrthographicCameras) → torch.Tensor[source]#

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