Jesse Grabowski

Add Ops for LU Factorization

2

## Description This PR will add the following Ops: - [x] lu - [x] lu_factor - [x] lu_solve As well as dispatches for numba/jax (and maybe torch, though help is...

Add Ops for specialized dot products with structured matrices

3

### Description BLAS/LAPACK offer a number of specialized dot products when matrices have special structure: - Triangular [matrix-vector product ](https://netlib.org/lapack/explore-html/d6/d1c/group__trmv_ga73370bd6dca01abe05d54ecd1d91ce9a.html#ga73370bd6dca01abe05d54ecd1d91ce9a) and [matrix-matrix product](https://netlib.org/lapack/explore-html/dd/dab/group__trmm_ga4d2f76d6726f53c69031a2fe7f999add.html#ga4d2f76d6726f53c69031a2fe7f999add) - Symmetrical [matrix-vector product ](https://netlib.org/lapack/explore-html/d0/d4b/group__hpmv_ga739f8dc2316523832bde2b237fcad8a6.html#ga739f8dc2316523832bde2b237fcad8a6) and [matrix-matrix...

Replace our `Tri` op with an `OpFromGraph`

1

### Description Currently we have an Op that calls `np.tri`, but we can very easily build lower triangular mask matrices with `_iota`: ```py from pytensor.tensor.einsum import _iota def tri(M, N,...

### Description Right now we assume the inputs to `solve_triangular` are always consistent with the settings passed to the function. For example, if one sets `lower`, we assume that the...

Add `pytensor.tensor.optimize`

8

## Description Implement scipy optimization routines, with implicit gradients. This PR should add: - [ ] `optimize.minimize` - [ ] `optimize.root` - [ ] `optimize.scalar_minimize` - [ ] `optimize.scalar_root` It...

Add numba overload for `scipy.linalg.svd`

2

### Description We currently use `np.linalg.svd`, which doesn't support `compute_uv = True`, or inplace flags. The relevant LAPACK functions to get full functionality are [gesdd ](https://netlib.org/lapack/explore-html/df/d22/group__gesdd.html) and [gesvd](https://netlib.org/lapack/explore-html/d1/d7f/group__gesvd.html).

Add commonly used priors over PSD matrices

2

## Tell us about it I'm thinking specifically about LKJCholeskyCov and/or LKJCholeskyCorr, but also Wishart/InverseWishart. I use these often enough but don't have strong intuitions about them. Would be very...

Support for distance metrics

2

Would distance metrics like wasserstein and kl-divergence (ok ok it's not a proper distance) be within the scope of this project? It would be nice to have an out-of-the-box way...

**Describe the bug** For either float32 or float64 inputs, `mx.linalg.cholesky` seems to occasionally disagree with the LAPACK implementation used by scipy. Highly possible I'm missing about doing full-precision computation --...

1. This should be shown when you just look at the *model*, without having to call a separate function on the GCN file 2. It should include summary information about...