Media creation with artificial intelligence: Difference between revisions

This would include to train own models. The idea is to let the models have a neural representation of objects that equals to screenshot-taking from 3D so that prompts will output almost never hallucinate details. That way artist can reduce post-editing as the generated outputs also include his own style.

Own models once more boost rapid prototyping because they reduce the necessity to have a more complexer, combined 3D-2D-workflow.