-
Suggestions for improving the mail analysis feature:
1. Allow customization of `html-validate` settings in `appsettings.json`. Current rules aren't always suitable for email HTML.
2. Add option …
-
It is my understanding that triton does not yet support complex datatypes, so the only way is to probably represent complex numbers with `tl.float32` tensors with an extra dimension of 2 (for the real…
-
Complex numbers would have ffi.complex set as metatable (no modifications allowed as per usual FFI ctypes).
It would be written in Lua, with some C calls for the tricky functions. If this is the case…
-
In Python at least, `np.std` on complex numbers yields NaN, as discussed in #3420:
![image](https://github.com/posit-dev/positron/assets/329591/2a647abd-9f4d-4335-91e0-2334fa7d4de2)
We should de…
wesm updated
3 months ago
-
## About the author
I am Endre, a data scientist in academia - sort of a research software engineer. Most of my current work deals with statistics, programming, complexity and scaling but I think I…
-
Since #3 is too general, I am opening this to track the changes.
-
```
'(1+1j) * dw_xxx...' results in parse error.
the only way to do this is '1 * dw_xxx.. + 1j * dw_xxx...', but it is not
efficient.
```
Original issue reported on code.google.com by `vlukes@kme.zc…
-
Something like:
```
VectorXcd b;
b
-
NCCL does not support complex numbers directly and does not plan to ([see issue](https://github.com/NVIDIA/nccl/issues/539)). Are we willing to add a wrapper to NCCL.jl to make using complex numbers …
-
### Problem Description
The feature of automatic scaling for formations (as was implemented in #4471) was much appreciated, but the (MultiAxisPoint) implementation in PR 4471 was very complex; more c…