J.Y.

icon location This GH account is used for work-related contributions to open source software projects. See @bobye for my other personal projects.

Results 149 comments of J.Y.

With large dataset, Splatfacto never densify gaussians

We need better mechanism of resetting and splitting to handle large set of input images.

[gsplat] on Windows, spherical harmonics consume excessive VRAM

This issue might be fixed in the latest PR: https://github.com/nerfstudio-project/nerfstudio/pull/2709 I have successfully trained GS on bicycle dataset.

'pixel_sampler' not defined in exporing point clouds with 'vanilla-nerf' and 'splatfacto'

not all models support point cloud exports.

Fix SH

Just to cite Inria gaussian splatting's implenentation: https://github.com/graphdeco-inria/gaussian-splatting/blob/main/utils/sh_utils.py

Fix SH

Just want to revisit this topic: Shall we remove SH evaluation in nerfstudio, and use the [one](https://github.com/nerfstudio-project/gsplat/blob/main/gsplat/_torch_impl.py#L12) from gsplat library? So it minimize our maintenance effort. @kerrj

AssertionError: block_width must be between 2 and 16 in the latest version 1.0.3.

Could you update the gsplat==0.1.10? `pip install --upgrade gsplat`

AssertionError: block_width must be between 2 and 16 in the latest version 1.0.3.

This issue seems to be a dup of https://github.com/nerfstudio-project/nerfstudio/issues/3033 But I am not why it happens with nerfstudio=1.0.3 + gsplat==0.1.10 I cannot reproduce the issue locally.

Support OCI standard dockerfile

@Zunhammer Could you follow up this PR? Does the current state look good to you?

splatfacto: "RuntimeError: can't retain_grad on Tensor that has requires_grad=False"

Do you have sample datasets that we can reproduce the error?

Jkulhanek/zipnerf

Close as dup of https://github.com/nerfstudio-project/nerfstudio/pull/2850