Remaining issues

[constantinpape]

Here's a list of remaining issues with the functionality here I found while working on #92.
It would be good to fix some of this before making a release, maybe except for last two points, which could be a bit more complicated.

• prediction_pipeline (and transitively also prediction functions) only support a single input/output tensor; need example model for this
• prediction_pipeline misses output_dtype
• prediction_pipeline does not support models with fixed output shape
• the pre/postprocessing is not quite consistent with https://github.com/bioimage-io/spec-bioimage-io/blob/main/supported_formats_and_operations.md
• prediction with tiling currently only works for same output shape, but supporting changes in output shape is possible (need example model)
• issues with normalization only based on local patches, e.g. for tiling, should fix by implementing and using the per_dataset mode

[constantinpape]

@FynnBe I think it would be good to work on this (at least first 4 points) before releasing this library. I can work on the pre/post-processing things next. And we can maybe discuss how to split the other things in the bioimage.io meeting.

[constantinpape]

Support for multiple tensors has been implemented in #103 and #104 and the output_dtype is not relevant any more since prediction_pipeline has output_specs now.

@FynnBe do you want to tackle fixed shapes next? You can use this model for tests.

Also, do you want to work on removing the attributes that duplicate stuff from input/output spec in prediction_pipeline? I think it would be good to also simplify that part before the release.

[FynnBe]

Yes and yes 👍

[FynnBe]

Also, do you want to work on removing the attributes that duplicate stuff from input/output spec in prediction_pipeline? I think it would be good to also simplify that part before the release.

simplified in #106

we could take it a step further and remove crreate_prepdiction_pipeline and use its __init__ instead, as it's rather trivial now.

[FynnBe]

@FynnBe do you want to tackle fixed shapes next? You can use this model for tests.

fixed shapes PR: #105