"The Caption Captain"
A simple, beginner friendly tool with a guided approach to providing detailed image captions for your training data set.
Anyone that has ever created an image data set to train a model would agree, captioning your images is a drag. It's time consuming and monotonous, even for the most experienced model makers.
Now consider the users that are just starting to dabble in the world of generative AI. Captioning a dataset can be overwhelming and intimidating.
Where do I start?
How much detail to go into?
How should I word the tags?
Is there anything I don't want to tag?
What's the right way I'm suppose to tag my images?
These are all questions that someone who has never trained a LoRA dataset might have. This is where The Caption Captain comes in handy. I built this CustomGPT for help with tagging my own data set images as i quickly discovered that most of the auto-captioning functions don't do as well with photorealism, nor do they provide then in a format that makes it easy to prompt. That's not even including the mislabel, duplicates and the fact that they often end up tagging details related to the concept you're trying to train which just confuses things.
