kaolin-wisp
kaolin-wisp copied to clipboard
NVIDIA Kaolin Wisp is a PyTorch library powered by NVIDIA Kaolin Core to work with neural fields (including NeRFs, NGLOD, instant-ngp and VQAD).
I'm trying to build a generalizable NeRF [e.g., pixelNeRF] where the features are not scene specific but taken from some encoder. I'm trying to use an SPC Field with an...
There a couple files missing a typedef for uint which is causing a compilation error grid_interpolate_cuda.cu uniform_sample_cuda.cu
## Description Hi, I tried to run main_nerf.py in the main branch. But it suddenly stopped showing a one-word line `Killed`. It is presumably due to RAM shortage, according to...
Hi there. First of all, thank you for open sourcing this super useful repo. I wanted to do pose optimization within a wisp pipeline, leveraging the `kaolin.Camera` class, which is...
@ttsesm You can add multiple objects to the scene graph like this: ```python3 from wisp.framework import WispState from wisp.renderer.core.api import add_to_scene_graph wisp_state = WispState() # Your global shared info here.....
When I try to run main_nerf.py, I successfully loaded the data, but the interactive renderer was crushed due to the following errors: ``` 2023-05-11 20:07:33,458| INFO| [i] Running at 60...
The current `PackedRFTracer` implementation allows some control for `bg_color` to the RGB channel. We're missing this flexibility for additional channels (i.e. whose "clear color" may be random or some other...
Subclassing `WispDataset` to, i.e, return a `MultiviewBatch` with additional supervision fields, and applying `SampleRays` transform on top may cause some of the fields to be omitted. For example, in the...
https://github.com/NVIDIAGameWorks/kaolin-wisp/blob/cb47e10f376e5ac8b6965c650d8a6b85b9bc968e/wisp/tracers/packed_rf_tracer.py#L148-L153 I have a doubt in this section of the code. When the background is white what is the interpretation of adding "ray_colors" to "(1-alpha)"? Will this not make the...
In addition to the current SPC raymarcher, the following NGP features would be useful to integrate: 1. Bottom level acceleration structure: occupancy bitgrid 2. DDA ray marcher, backed by a...