santa hat
deerdeer nosedeer glow
Sign In
Model testing card

A few days ago I stumbled upon a reddit post with a model testing card https://www.reddit.com/r/StableDiffusion/comments/1b39b07/model_testing_card_idea/.

I decided to improve this a bit with centrally controlled parameters for the generation and support for LoRA and Lightning LoRA.

Some things could probably be improved like better prompts for photo/anime comparison as I don't get too much variety with LoRA enabled.

All the input parameters are on the left side with the final output (as a preview that can be saved) there too.

Updates:

  • v3: use LoRA Auto Trigger so that the trigger word doesn't need to be added manually

  • v2: Average conditionings for Portrait and Closeups instead of concatenating prompts

Models tested:

LoRAs tested:

130

Comments