core-bioimage-io-python
core-bioimage-io-python copied to clipboard
Python libraries for loading, running and packaging bioimage.io models
For some reason, some weights files get corrupted after downloading, and it get saved in the cache so it always crash no matter how many times I download it. I...
It seems when calling `export_resource_package` it always create a zip package but it takes significant more time to load the zip package for inference compared to an extracted directory. It...
I am trying to maintain a installed models folder by using the following `download_model` function, then use `bioimageio.core.load_resource_description(model_path)` to load the function, but it failed for this model: "10.5281/zenodo.5911832", it...
Would it make sense to somehow link the bioimageio.spec versions with the core, namely, when running `pip install -U bioimage.core` I am kind of hope it will also upgrade the...
I have been working on the widefield models to upload them to the zoo, but there is an issue with the test validator. The models have inputs of 3 dimensions...
Currently, the pytest fixtures for getting single models need to be generated by hand here: https://github.com/bioimage-io/core-bioimage-io-python/blob/main/tests/conftest.py#L177-L229. It would be nice to generate these automatically instead.
@FynnBe how do we currently deal with `per_dataset` normalization? I could not find this documented anywhere. And if I run prediction with a `per_dataset` model naively I get the following...
I am trying to run model testing from a doi. In the docs it said I should do `bioimageio test-model -m ` But it said `-m` is not an option:...
> Slightly unrelated: one of the issues with the marshmallow error messages is that it does not tell you which spec it is comparing to. E.g. when you use `jsonschema`...
> weirdly it looks like the test input/output files do not exist? Which might also explain the error if instead of the .npy file we download some 404 response. I've...