Anthony Zhang's Avatar

Anthony Zhang

@antoniomacaronio.bsky.social

Nerfstudio contributor, fan of fast software 😎

35 Followers  |  152 Following  |  9 Posts  |  Joined: 22.10.2024
Posts Following

Posts by Anthony Zhang (@antoniomacaronio.bsky.social)

Video thumbnail

My entry for the @dev.to CSS Art challenge is taking shape...

24.07.2025 12:14 β€” πŸ‘ 64    πŸ” 6    πŸ’¬ 11    πŸ“Œ 0

I agree - setting up an environment is so painful and while claude/gpt/etc are okay at helping dependency issues if codex can automate this it will be so nice

17.05.2025 17:49 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

A thread of thoughts on radiance fields, from my keynote at 3DV:

Radiance fields have had 3 distinct generations. First was NeRF: just posenc and a tiny MLP. This was slow to train but worked really well, and it was unusually compressed --- The NeRF was smaller than the images.

08.04.2025 17:25 β€” πŸ‘ 93    πŸ” 21    πŸ’¬ 2    πŸ“Œ 1

looks very cool! am excited to compare this to megasam

02.04.2025 02:39 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

okay I fell for this and now I’m saving this monumental piece of work 😭😭😭☠️☠️☠️

01.04.2025 14:47 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

ahh this is my mistake - it is just image, pointmap datapoints

I am still impressed by sora’s ability to learn 3D without 3d info (like pointmaps)!

11.03.2025 14:10 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

i’m not gonna lie the naming is hilarious 😭😭☠️ alligator va croco

11.03.2025 13:49 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

post submission cycling is like mcdonalds sprite after a hangover

11.03.2025 13:44 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

I’m honestly still shocked at this - how did Sora manage to learn 3D consistency just from videos?

Dust3r (i know it’s completely different) had to be given camera parameters during its training process, so to me it makes sense it was able to understand 3D very well

11.03.2025 13:42 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 3    πŸ“Œ 0
Preview
Splatfacto Paper Website 3D Gaussian Splatting was proposed in SIGGRAPH 2023 from INRIA, and is a completely different method of representing radiance fields by explicitly storing a collection of 3D volumetri...

if you have a CUDA machine, you can try nerfstudio’s splatfacto method if you want a video of the optimization process on custom data

there are some good videos on the nerfstudio documentation website too!

docs.nerf.studio/nerfology/me...

11.03.2025 13:39 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
(🎁) Add a command to "upgrade" dependency constraints in the `pyproject.toml` · Issue #6794 · astral-sh/uv this will update the version pins in the pyproject.toml file prior art: poetry: https://github.com/MousaZeidBaker/poetry-plugin-up npm: https://docs.npmjs.com/cli/v10/commands/npm-update > uv upgra...

Gave #uv a try for managing my #python dependencies. This missing feature is quite important for my flow hence a deal breaker:

github.com/astral-sh/uv...

Following the GitHub issue to revisit when resolved but till then #nix + #poetry > #uv

17.02.2025 10:06 β€” πŸ‘ 3    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0

the way cameraHMR managed to get the hands flat for the handstand picture was very neat!!

30.01.2025 19:31 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0