First Impressions of Leonardo AI - Using Different Models

Leonardo AI allows you to generate images from a braod range of models, some of which are set in the default list, others are found in the Community list of models, and then there are even more models that can be found by seaching for a term. In this case the final models were searhed under the term "Fantasy".

The prompt used for all but one generation (as explained below) was the same: "photorealistic image of a wood elf emerging from a hollow in a tree, background is an old growth forest, natural lighting at dawn".

The results varied greatly in quality. My general impression was that there were issues with the number of limbs, misidentification of arms for legs and vice-versa, and the usual problems with numbers of digits, as well as the shape and structure of eyes.

Resolution was also an issue with some of the model where they were trained at 512x512 px. The max resolution was 1024x1024 Px.

Of the models sampled, the ones I would plan to use again are Leonardo Diffusion, Character Portrait and RPG 4.0.

- Rob Fossett, March 17, 2023.

You are welcome to copy any of the images for personal use. Commercial use is prohibited, and this includes resales within Second Life. The AI graphic artists who created the images hold copyright (C) over their works.

NSFW images are not permitted on this website.