Vickie Ye

researcher @AnthropicAI. previously ML for 3D and graphics @berkeley_ai PhD, BS @MIT EECS + physics

[Anthropic @AnthropicAI]

Our interpretability team recently released research that traced the thoughts of a large language model.

Now we’re open-sourcing the method. Researchers can generate “attribution graphs” like those in our study, and explore them interactively.

[Zhengqi Li @zhengqi_li]

Introducing MegaSaM! 🎥

Accurate, fast, & robust structure + camera estimation from casual monocular videos of dynamic scenes!

MegaSaM outputs camera parameters and consistent video depth, scaling to long videos with unconstrained camera paths and complex scene dynamics!

[Vickie Ye @vickie_ye_]

@JonathonLuiten Thanks!

[Vickie Ye @vickie_ye_]

Amazing work Ruilong!! So excited to continue to build gsplat with you!

[Ruilong Li @ruilong_li]

Excited to announce 🚀gsplat v1.0🚀: a ⏩efficient⏩ CUDA backend for 3D Gaussian Splatting!
https://t.co/2nREWuEYwX
A drop-in replacement of the official impl. with:
- Up to 2x faster training;
- Up to 4x less GPU memory;
- Render millions of GSs in real-time;
- And more;

[Vickie Ye @vickie_ye_]

This is an amazing visual walkthrough of SLAHMR, @leonard_bruns! Looking forward to the PR with @rerundotio integrated!

[Leonard Bruns @leonard_bruns]

What SLAHMR does is super impressive! It robustly tracks the motion of people filmed with a moving camera. It’s a great showcase of how to build robust CV systems by combining several single purpose models. I used @rerundotio to dig into the implementation and it’s pretty cool.

[Leonard Bruns @leonard_bruns]

What SLAHMR does is super impressive! It robustly tracks the motion of people filmed with a moving camera. It’s a great showcase of how to build robust CV systems by combining several single purpose models. I used @rerundotio to dig into the implementation and it’s pretty cool.

[Vickie Ye @vickie_ye_]

@_akhaliq thanks for bringing this up! I've updated it -- let me know if you still have issues