-
### Describe the bug
I am using an InvariantMetric equipped on SO3 to compute a geodesic between two points. Besides not being able to equip the metric using the function `geometry.manifold.Manifol…
-
Let's try to add some diagnostic functionality to AIBECS that generalizes the work done in *Pasquier and Holzer*, 2018.
---
### Preliminary notes
Let's start from the generic equation
(∂ₜ …
-
I tested Tacho (fine grained task cholesky code) using the Kokkos cuda task dag capability. With help from @crtrott , we can profile the performance of the task dag.
With a enough number of concur…
-
Use Diagonalwise conv instead of Depthwise Conv, or use TensorFlow Depthwise Conv implementation.
Diagonalwise Refactorization: An Efficient Training Method for Depthwise Convolutions
* https://…
-
This is an option that can potentially have much better performance compared to using the dense cublas-based solver. It should be compared to the cuSparse based solver.
The factorization itself doe…
-
I've noticed that MatX needs to include some CUDA-related abstractions and facilities (like error handling) which are actually not specific to matrix/tensor/numeric computations in any way.
I woul…
-
Hello,
I recently started using rsatoolbox in python in order to compute Mahalanobis distances (and eventually crossnobis + taking noise variance into account). I have an unbalanced design - the n…
ndolf updated
7 months ago
-
In Section 3.4.1 (Computational Requirements: Memory) of your paper "PRIMME: PReconditioned Iterative MultiMethod Eigensolver: Methods and software description" it is mentioned that the amount of allo…
-
## 🚀 Feature
Implement the positive-definite-matrix-valued Wishart and inverse-Wishart distributions, currently missing from `torch.distributions`.
## Motivation
These distributions often pop…
-
The following problem has been observed by Hongyi Zhao.
It happens both in GAP 4.12.2 and in the master branch.
```
gap> G1:= AffineCrystGroupOnLeft(
> [ [ [ -1, 0, 0, 0 ], [ 0, 1, 0, 1/4 ], …