Artur Toshev

Happy to introduce 🔥LaM-SLidE🔥!

We show how trajectories of spatial dynamical systems can be modeled in latent space by

--> leveraging IDENTIFIERS.

📚Paper: arxiv.org/abs/2502.12128
💻Code: github.com/ml-jku/LaM-S...
📝Blog: ml-jku.github.io/LaM-SLidE/
1/n

22.05.2025 12:24 — 👍 7    🔁 8    💬 1    📌 1

The Gaussian is a nice bumpy shape, but sometimes we hope for a smooth (i.e. C∞) function like the Gaussian that is 𝒂𝒍𝒔𝒐 compactly supported.

One such class of functions is called "Bump functions"

1/6

29.11.2024 21:38 — 👍 139    🔁 19    💬 4    📌 2

Accelerating MACE at inference by 17x sounds insane. But I'm curious to see the numbers other ppl get when they replace e3nn in their codes. Also, not sure about the speedups on hardware older than H100/A100.

21.11.2024 22:11 — 👍 0    🔁 0    💬 0    📌 0

Accelerate Drug and Material Discovery with New Math Library NVIDIA cuEquivariance | NVIDIA Technical Blog AI models for science are often trained to make predictions about the workings of nature, such as predicting the structure of a biomolecule or the properties of a new solid that can become the next…

sorry, I forgot to attach the blog: developer.nvidia.com/blog/acceler...
But I don't know more than that

21.11.2024 21:58 — 👍 1    🔁 0    💬 1    📌 0

GitHub - NVIDIA/cuEquivariance: cuEquivariance is a math library that is a collective of low-level primitives and tensor ops to accelerate widely-used models, like DiffDock, MACE, Allegro and NEQUIP, ... cuEquivariance is a math library that is a collective of low-level primitives and tensor ops to accelerate widely-used models, like DiffDock, MACE, Allegro and NEQUIP, based on equivariant neural n...

⚛️ cuEquivariance ⚛️ is something I've been waiting for forever. Thank you Mario Geiger and team!

Curious to see where the "equivariance is just too slow" debate goes 👀

github.com/nvidia/cuequ...

21.11.2024 21:47 — 👍 10    🔁 0    💬 1    📌 2