LoRa description

These LoRa's can be called as "Photo Moodificators". Made from travel photographs. No particular theme. Just a random moody photos from travels (nature, streets, animals, people, etc.).

Prompting

I've used a mixed approach to caption the images in dataset. The captions in Training Data (captions.zip) contains the mix of WD tags and natural description made by LLM 7b (Llama Mistral 7b).

So, the best advice on getting idea about prompt words is to look into the captions and the example images I am posting here.

Part 1 - first 500 captions.

Part 2 - next 500 captions.

Part 3 - last 200-300 captions.

PS: sorry... no images in captions.zip. I feel very shy about some of my photographs.