-
Hello,
I've been playing around with AutoDiff and I've been able to make the reverse mode differentiation work with some code I have. However I was wondering if it is possible to traverse the expre…
-
## Description
A few questions have come up on the forums lately regarding integrate_1d where the integrator fails and it is difficult to debug what is going on.
Specifically I am talking about:…
-
## The issue:
Doing `jit`ted `scan`s with variable `length`s throws `jax.core.ConcretizationTypeError` as evidenced in the following minimal example
## Minimal example:
Suppose we want to solve a…
-
Even though `stack` and `concatenate` are nice for combining arrays, sometimes they don't fit the data I have or require a significant amount of work to use. For instance, combining blocks of differen…
-
I am loving JAX, and coding with it has been amazing.
I am encountering unexpected behavior when I try to calculate the gradient of integrals which contain singularities, even when I replace these…
smiet updated
3 years ago
-
## Description
@SteveBronder found that softmax isn't quite working right. The reverse mode softmax code here sets the adjoint instead of incrementing it: https://github.com/stan-dev/math/blob/deve…
-
#### Summary:
This is to track the progress of the timing/profiling project.
A prototype is implemented following the design [here](https://discourse.mc-stan.org/t/proposal-for-profiling-stan-mo…
-
Calling `autodiff` on `MeasureTheory.logdensity` for a multivariate distribution causes Enzyme (and Julia) to crash. Here is a MWE (open details to see the huge stacktrace):
```julia
julia> using …
-
## Description
SVD(X) = U * Sigma * V^T
We currently have a function for the singular values (the diagonal of Sigma): https://mc-stan.org/docs/2_25/functions-reference/linear-algebra-functions-a…
-
Hi @xukai92,
As discussed on Slack, adding support for complex numbers would allow us to use many models from physics. The code below is a quantum model of human judgment based on Wang et al. (201…