Ziwen Liu
Ziwen Liu
Late to the discussion, but happy to move [this code](https://github.com/czbiohub/iohub/blob/eeaad0861bf92e53a8b931596fd4002ddd7d5860/iohub/ngff_meta.py) upstream for v0.4 too!
> This approach keeps all the pydantic models separate from the ome-zarr-py classes. > > But I'm wondering whether we should be thinking about combining them. E.g. If combining them...
On this note, since we are having many more models, the old initialization scheme based on dictionaries is starting to be restricting. I have been thinking about using [dependency injection...
We also potentially need better names to distinguish between: - 3D stem -> 2D encoder -> 2D decoder -> ***2D*** head (currently called 2.1D) - 3D stem -> 2D encoder...
> 2.5D UNeXt: 2.5D, but with ConvNeXt layers > 3D UNeXt: 3D, but with ConvNeXt layers Can we call these nD *L*UNeXt for consistency?
Thanks @normanrz! I endorse this RFC.
Got segfault too: log ```journal kernel: python[28489]: segfault at 7f5300a5cc60 ip 00007f5112633249 sp 00007ffd80ffe1d8 error 4 in _pyllamacpp.cpython-310-x86_64-linux-gnu.so[7f51125e1000+64000] likely on> kernel: Code: 00 48 89 c2 83 e0 01 48...
It's a temporary workaround for #164. If you unpin it and tests pass it's fine to revert.
@AhmetCanSolak Thanks for moving this discussion over. I also mentioned the possibility of something like `skimage.io.imread` during an offline meeting. And I think this functional entry point may alleviate some...
I've started working on the `imread` idea in #71. Still WIP but early-stage input is very welcome!