Commit 7d8d8eb4 authored by Boris Bonev's avatar Boris Bonev Committed by Boris Bonev
Browse files

minor cleanup

parent 6a845fd3
...@@ -3,10 +3,10 @@ The code was authored by the following people: ...@@ -3,10 +3,10 @@ The code was authored by the following people:
Boris Bonev - NVIDIA Corporation Boris Bonev - NVIDIA Corporation
Thorsten Kurth - NVIDIA Corporation Thorsten Kurth - NVIDIA Corporation
Max Rietmann - NVIDIA Corporation Max Rietmann - NVIDIA Corporation
Mauro Bisson - NVIDIA Corporation
Andrea Paris - NVIDIA Corporation Andrea Paris - NVIDIA Corporation
Alberto Carpentieri - NVIDIA Corporation Alberto Carpentieri - NVIDIA Corporation
Mauro Bisson - NVIDIA Corporation
Massimiliano Fatica - NVIDIA Corporation Massimiliano Fatica - NVIDIA Corporation
Christian Hundt - NVIDIA Corporation
Jean Kossaifi - NVIDIA Corporation Jean Kossaifi - NVIDIA Corporation
Nikola Kovachki - NVIDIA Corporation Nikola Kovachki - NVIDIA Corporation
Christian Hundt - NVIDIA Corporation
\ No newline at end of file
...@@ -11,7 +11,7 @@ ...@@ -11,7 +11,7 @@
* S2 Transformer * S2 Transformer
* S2 Segformer * S2 Segformer
* S2 U-Net * S2 U-Net
* Reworked spherical examples for better reproducibility * Reorganized examples folder, including new examples based on the 2d3ds dataset
* Added spherical loss functions to examples * Added spherical loss functions to examples
* Added plotting module * Added plotting module
......
...@@ -171,7 +171,7 @@ Here, $x_j \in [-1,1]$ are the quadrature nodes with the respective quadrature w ...@@ -171,7 +171,7 @@ Here, $x_j \in [-1,1]$ are the quadrature nodes with the respective quadrature w
### Discrete-continuous convolutions on the sphere ### Discrete-continuous convolutions on the sphere
torch-harmonics now provides local discrete-continuous (DISCO) convolutions as outlined in [4] on the sphere. These are use in local neural operators to generalize convolutions to structured and unstructured meshes on the sphere. torch-harmonics now provides local discrete-continuous (DISCO) convolutions as outlined in [5] on the sphere. These are use in local neural operators [2] to generalize convolutions to structured and unstructured meshes on the sphere.
### Spherical (neighborhood) attention ### Spherical (neighborhood) attention
...@@ -276,14 +276,22 @@ Spherical Fourier Neural Operators: Learning Stable Dynamics on the Sphere; ...@@ -276,14 +276,22 @@ Spherical Fourier Neural Operators: Learning Stable Dynamics on the Sphere;
International Conference on Machine Learning, 2023. [arxiv link](https://arxiv.org/abs/2306.03838) International Conference on Machine Learning, 2023. [arxiv link](https://arxiv.org/abs/2306.03838)
<a id="1">[2]</a> <a id="1">[2]</a>
Liu-Schiaffini M., Berner J., Bonev B., Kurth T., Azizzadenesheli K., Anandkumar A.;
Neural Operators with Localized Integral and Differential Kernels;
International Conference on Machine Learning, 2024. [arxiv link](https://arxiv.org/abs/2402.16845)
<a id="1">[3]</a>
Schaeffer N.; Schaeffer N.;
Efficient spherical harmonic transforms aimed at pseudospectral numerical simulations; Efficient spherical harmonic transforms aimed at pseudospectral numerical simulations;
G3: Geochemistry, Geophysics, Geosystems, 2013. G3: Geochemistry, Geophysics, Geosystems, 2013.
<a id="1">[3]</a> <a id="1">[4]</a>
Wang B., Wang L., Xie Z.; Wang B., Wang L., Xie Z.;
Accurate calculation of spherical and vector spherical harmonic expansions via spectral element grids; Accurate calculation of spherical and vector spherical harmonic expansions via spectral element grids;
Adv Comput Math, 2018. Adv Comput Math, 2018.
<a id="1">[4]</a> <a id="1">[5]</a>
Ocampo, Price, McEwen, Scalable and equivariant spherical CNNs by discrete-continuous (DISCO) convolutions, ICLR (2023), arXiv:2209.13603 Ocampo, Price, McEwen, Scalable and equivariant spherical CNNs by discrete-continuous (DISCO) convolutions, ICLR (2023), arXiv:2209.13603
<a id="1">[6]</a>
Bonev B., Rietmann M., Paris A., Carpentieri A., Kurth T.; Attention on the Sphere; [arxiv link](https://arxiv.org/abs/2505.11157)
\ No newline at end of file
...@@ -60,3 +60,6 @@ dev = [ ...@@ -60,3 +60,6 @@ dev = [
"PIL", "PIL",
"h5py", "h5py",
] ]
[tool.black]
line-length = 180
\ No newline at end of file
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment