Closed ClementCaporal closed 2 months ago
Hello thank you for your review @danifranco I changed the structure as requested. Tell me if I can do something else to match BiaPy pattern.
Just few minor things more:
DATA.PREPROCESS.ZOOM.ZOOM_FACTOR
above description please. Like: "E.g. [1,2,2,2,3] that needs to match Zarr axes"? or something like that. TEST.BY_CHUNKS.INPUT_IMG_AXES_ORDER
or TEST.BY_CHUNKS.INPUT_MASK_AXES_ORDER
? If so put a check in check_configuration.py please.Hello !
Hello! Thank you so much for doing these changes Clement! I have a suggestion here:
_C.DATA.PREPROCESS.ZOOM
to follow the same preprocessing convention we are already using. Even if the zoom can only be aplied when predictingby_chunks
I would do that way. I'd put in it's comment/description above a message warning the user that is only enabled in that case (see other cases inconfig.py
that are described like that so you can follow the same message structure in case you have doubts). E.g. for Canny preprocessing this is hwo it's done: https://github.com/BiaPyX/BiaPy/blob/f872e1fd74fe55b66c694ef0289bb07495265c5f/biapy/config/config.py#L320