-
### Description:
CPU memory is not released when sequentially training independent neural networks that output a distributions.multivariate_normal.MultivariateNormal object. This is in contrast to …
-
Discussed elsewhere, opening an issue to track it.
Since they have a slightly different interface and could benefit from different optimizations, @andreasnoack has suggested we separate multivariat…
-
## Description
Looking at the [Stan docs](https://mc-stan.org/docs/2_27/functions-reference/multivariate-normal-distribution.html) we only accept arrays of vectors and row vectors for the multiva…
-
Many empirical calibrations of parameters use multivariate Gaussians to represent the covariance between parameters. For example, a quadratic damage function with two coefficients, alpha and beta, wou…
-
We support multivariate distributions and nalgebra is part of our API. We'll be merging #209 and this will be a breaking change. The `new` methods clearly rely on `nalgebra`, but it might be more impl…
-
This issue is motivated by ongoing work in SciPy to improve the existing multivariate distributions and add new ones. We'd like to be sure that, if possible, enhancements that we make remain consiste…
-
### 🐛 Describe the bug
I'm trying to recreate tensorflow's `MultivariateNormalDiagWithSoftplusScale` so I went ahead and did this
```
dist = MultivariateNormal(mu.squeeze(), torch.diag_embed(F.sof…
-
Please consider adding the complex multivariate normal distribution.
A helpful reference: [Picinbono, B. (1996). Second-order complex random vectors and normal distributions.](http://ai2-s2-pdfs.s3…
-
Thanks for creating such a nice library.
I have experienced numerical instability in the Multivariate Gaussian Distributions. For better numerical stability, I wondered if it would be better replac…
-
*Note from @mdhaber: This is a deep issue that is slated to be addressed as part of [this project](https://chanzuckerberg.com/eoss/proposals/scipy-fundamental-tools-for-biomedical-research/). Before c…