I organize my images with different tags to describe the scenario of which each image was taken and its notable attribute. But doing so manually slow down my flow.
In the best case, the data I got is already sorted and separated to my liking and can be batch tagging during upload. But in most case it doesn’t, especially public datasets.
My solution is to train a classification model, run it through the dataset and binning them to corresponding folder.
I would love to see this get implemented in your tool. It should be similar to the use of custom train model to help labeling image, just now on the tag instead.
My apology if this feature is already available, suggested or down right absurd.
Also feel free to correct my broken English, appreciate.