Pick a short and memorable name, like pixora-model. You can use lower case characters, digits, periods, underscores and dashes.
A zip file containing the images that will be used for training. We recommend a minimum of 10 images. If you include captions, include them as one .txt file per image, e.g. my-photo.jpg should have a caption file named my-photo.txt. If you don't include captions, you can use autocaptioning (enabled by default).
The trigger word refers to the object, style or concept you are training on. Pick a string that isn’t a real word, like TOK or something related to what’s being trained, like CYBRPNK. The trigger word you specify here will be associated with all images during training. Then when you use your LoRA, you can include the trigger word in prompts to help activate the LoRA.Default: "TOK"
Automatically caption images using Llava v1.5 13BDefault: true
Optional: Text you want to appear at the beginning of all your generated captions; for example, 'a photo of TOK,'. You can include your trigger word in the prefix. Prefixes help set the right context for your captions, and the captioner will use this prefix as context.
Optional: Text you want to appear at the end of all your generated captions; for example, 'in the style of TOK'. You can include your trigger word in suffixes. Suffixes help set the right concept for your captions, and the captioner will use this suffix as context.
Number of training steps. Recommended range is 500-4000.Default: 1000
Higher ranks take longer to train but can capture more complex features. Caption quality is more important for higher ranks.Default: 16