Élie Goudout

Thanks for your answer, I've stumbled upon this because while reading FGSM paper (the reference for implementation), I thought about generalizing to $L^p$ norms. Then I saw that this repo...

Actually, now that I think about it, I don't see any reason why this attack is not generalized to any $L^p$ noise. Let $p\in[1, +\infty]$ and $q$ such that $\frac{1}{p}+\frac{1}{q}=1$...

For example, suppose you have `fd` with `dim_domain = 3, dim_codomain = 4`. If I want to evaluate `fd` on an edge of its domain (which is a 2D surface...

The problem is fixed with #221, but not yet released or pushed on PyPi. You'll need to install from source for now: ```console git clone https://github.com/astanin/python-tabulate.git pip install ./python-tabulate python...

Just so you know, I got the [same issue](https://github.com/pytest-dev/pytest-timeout/issues/184) completely unrelated to `DeepSpeed` (I. didn't know the package until I came here). I think the culprit might be `pytest-timeout`, since...

Thank you, I did resort to unwrapping the model on my side. I don't think it's a bad approach because the final wrapper of Ultralytics is not meant for development...

I found a few references mentioning this, most notably [PEP 705](https://peps.python.org/pep-0705/#a-typedmapping-protocol-type) and a few related posts on the python forum: - https://discuss.python.org/t/pep-705-typedmapping/24827 - https://discuss.python.org/t/introduce-a-typedmapping-analog-to-typeddict-but-frozen/51905 One of the mentioned issue was...

I'm not entirely sure, but I think your approach forces inheritance from `TypedDict` which seems counterproductive (`TypedMapping` should be looser, am I right?)

Thanks again for your answers, I will close since you provided a quick workaround. May I ask though if/why you think this is not a good feature idea in principle?...

Thanks for the feedback 👍