Ludger Paehler

Oh, I see - I must have massively underestimated the difficulty here then..

If I may suggest, I would rather directly replace it with Pillow i.e. ```python from PIL import Image img = Image.open(file) img = np.array(img.resize((299, 299), resample=Image.BILINEAR)) ```

Shall we build this off of the paper examples?

Optimizations to be added to the `Using Enzyme` Subpage: - [ ] PhiOpt - [ ] SpecPhi - [ ] Mincut - [ ] Newcache LICM - [ ] AA

In addition, we should at the same consider supporting [hipStreams](https://rocmdocs.amd.com/en/latest/ROCm_API_References/HIP_API/Stream-Management.html) to maintain similar-ish support for both. Or are hip-streams easier to handle @vchuravy ?

Was just looking at the source of the particle framework our smoothed particle hydrodynamics codes here are built on top of, and differentiating through it would depend on our ability...

Coincidentally, Numba does actually seem to look ["good"](https://github.com/numba/numba/issues/7927) in that regard

Looking at the 6 test cases failing with LLVM-14, they should probably be fixed at some point but the error stays the same to the current errors on main.

Hey, did you manage to fix this by now? Would be great to have a little more information to help you: - What platform are you building on? - Against...

Would be great if you could bisect, when the error began appearing so we could fix it :)