Hashes the model files and records representation, channels, pixel size,
normalisation, license and the runtimes that can load the representation
(.pb for QuPath/OpenCV, SavedModel for QuPath/TensorFlow, bioimage.io
for the extension's bioimage.io support, a Python model directory for
stardist).
Usage
sg_stardist_manifest(
model,
...,
name = NULL,
channels = NULL,
n_channels_in = NULL,
pixel_size_um = NULL,
normalization = NULL,
license_id = NULL,
source = NULL
)Arguments
- model
Path to a
.pbfile, a SavedModel or Python model directory, a bioimage.io folder, or ansg_trained_modelwith backend"stardist".- ...
Reserved; must be empty.
- name
Model name (default: file/directory name).
- channels
Optional channel names the model expects.
- n_channels_in
Optional number of input channels.
- pixel_size_um
Optional pixel size the model was trained for.
- normalization
Optional list (e.g.
list(method = "percentile", low = 1, high = 99.8)).- license_id
Optional SPDX identifier.
- source
Optional short provenance text (no URLs are fetched).