Skip to content

Trying to quickly prototype data locally to apply the auto config option in the cloud

Jump to bottom
(CK) edited this page Aug 10, 2023 · 4 revisions

Figure 6: Run Tab

A passive part of the program logs captions that get edited to an auto-config.json file. This file gets edited everytime the user saves their progress editing tag/s and adding & removing images from their dataset. Tag/s removed from image/s are represented as well as tag/s added; tag/s added are given an insertion order such that if the config was ever applied, all changes would be in order. The real value of this feature comes in two forms.

  1. User can upload this auto-config.json file in the kilo-bytes size in comparison to entire datasets over 100+ images over an internet connection to train on the cloud and/or another computer. It's much faster.

  2. User can download their dataset in the cloud with very quickly and applying the auto-config.json file instantly applies all changes ever made to the dataset on the user's local machine. Given tag order is kept (by downloading) images in the same order.

To apply the changes, go to the Run Tab provide the path to the file and press the Apply Config Button

Clone this wiki locally