Adrian Seyboldt

Numba not vectorizing 2d copy loop

2

```python import numpy as np import numba import llvmlite.binding as llvm llvm.set_option('', '--debug-only=loop-vectorize,iv-descriptors') @numba.njit("void(float64[:, ::1], float64[:, ::1])", fastmath=True, boundscheck=False) def foo(a, out): n, m = a.shape assert out.shape == (n,...

Fix segfault for invalid axes in np.split

1

This adds a new function `normalize_axis` to `arrayobj.py`, and uses this function to verify axis arguments in `np.split`. There already is a similar function that could be wrapped in an...

Use of uninitialized values for empty loops

1

Surprisingly, this returns zero in numba ```python @numba.njit def foo(vals): for x in vals: pass return x foo(np.zeros(0)) ``` I think this is because it declares the loop variable before...

Segfault in np.split with negative axis

4

- [x] I have tried using the latest released version of Numba (most recent is visible in the change log (https://github.com/numba/numba/blob/main/CHANGE_LOG). - [x] I have included a self contained code...

Dims is not respected when broadcasting parameters

10

The dims argument is not respected when a random variable is created with parameters that broadcast to the desired output dimensions: ```python coords = { "dim1": pd.RangeIndex(10), "dim2": pd.RangeIndex(7) }...

pm.Model(coords=coords) should preserve coordinate type if it is a pandas Index

7

We currently convert all coordinates to tuples, but this removes valuable information, if the coordinates already are a pandas Index. I think we should only convert the input values if...

Add dict_to_array for logp_dlogp_function

3

Depending on what your PR does, here are a few things you might want to address in the description: + [ ] what are the (breaking) changes that this PR...

Trick to avoid python overhead

12

I figured out a way to avoid the python overhead if the right-hand-side function is implemented using numba, and I thought maybe people might be interested. The trick is to...

Handle duplicate indices in Numba implementation of `AdvancedIncSubtensor1`

5

Here are a few important guidelines and requirements to check before your PR can be merged: + [x] There is an informative high-level description of the changes. + [x] The...

Use objmode in scipy.special without numba-scipy

5

Here are a few important guidelines and requirements to check before your PR can be merged: + [x] There is an informative high-level description of the changes. + [ ]...