Vladimir Iglovikov
Vladimir Iglovikov
All transforms, except a few support any number of channels. We also check in tests for multiprocessing support. But yes, would be nice to add to docstrings.
What is the source of this code, i.e. what notebook?
Could you please a minimal example of the code that I could use to reproduce the issue?
@Andredance Thanks. I am rewriting the website right now. Should finish this week. Will look at your Notebook and add to docs.
Moved to https://github.com/albumentations-team/AlbumentationsX/issues/38
Will look into that.
Looks like the issue is addressed
Thanks. Will add `JpegCompression` to Benchmark and rerun with a new version of Kornia. https://github.com/albumentations-team/albumentations/?tab=readme-ov-file#benchmarking-results
```python Testing encoder: senet154 ❌ Failed: SENet.__init__() got an unexpected keyword argument 'strides' Testing encoder: se_resnet50 ❌ Failed: SENet.__init__() got an unexpected keyword argument 'strides' Testing encoder: se_resnet101 ❌ Failed:...
I am initializing as: ```python model = smp.Unet( encoder_name=encoder_name, encoder_weights=None, in_channels=1, classes=3, activation=None, ) ``` not passing strides as input