Ianna Osborne

Thanks, @nj-vs-vh ! Indeed, if `keepdims` is set to `True` the result looks correct: ```python >>> ak.mean(arr, weight=weight, axis=1, keepdims=True) ``` If `keepdims` set to `False` the result is as...

> > Thanks, @nj-vs-vh ! Indeed, if `keepdims` is set to `True` the result looks correct: > > ```python > > >>> ak.mean(arr, weight=weight, axis=1, keepdims=True) > > > >...

> PythonCall still seems to have trouble with Julia multi-threading (well, PyCall can have too, but it seems worse with PythonCall). So I'm not sure ... question is, do we...

```python def test_strings(): arr = ak.Array(["hey", "hi", "hum"]) > out = ak.to_cudf(arr) ^^^^^^^^^^^^^^^ arr = tests-cuda/test_3051_to_cuda.py:52: _ _ _ _ _ _ _ _ _ _ _ _ _ _...

> [@ianna](https://github.com/ianna) It should be an easy fix. It's just that the `StringColumn` constructor has changed: https://github.com/rapidsai/cudf/blob/d4961e1df7807fbacb1cfd49394cb4b8508d9a82/python/cudf/cudf/core/column/string.py#L92-L100 > > I just don't have a GPU atm to test :) I...

> @ianna some of the `tests-cuda-kernels-explicit` tests are failing. Do you know if this is expected? No, they all used to pass ... I'll have a look. Thanks!

> I'm not a big fan of this. We want the option to potentially dematerialize arrays to free up memory. We can now do so by clearing up the `_array`...

@jpivarski - yes, I can reproduce it. It did not fail when **firstly** the single test was run separately as: `python -m pytest tests-cuda/test_3149_complex_reducers.py ` but it did fail when...

fixed in https://github.com/scikit-hep/awkward/pull/3235

> This is waiting for a fix from Numba: [NVIDIA/numba-cuda#37](https://github.com/NVIDIA/numba-cuda/issues/37) and possibly on https://github.com/cupy/cupy/issues/8597