Trying the different samplers available in Stable Diffusion InvokeAI

A reference post

Trying the different samplers available in Stable Diffusion InvokeAI

This is just a reference post to share a comparison between the different samplers available. With different samplers you can get different output in your images.  I used InvokeAI's ability to use a text file to load up all the prompts and kick off a job for the MacBook Pro [see my last post on how to do that].

Each image used the same prompt, seed, cfg_scale, and steps. The only change was the sampler.

From my limited understanding, some of the samplers would benefit from a higher step count. k_dpm_2_a seems like it could use a few more to bake in as it looks blurry, and it is funny how different k_euler_a is from the rest.

Example prompt:

mdjrny-v4 style tiny cute cyber ninja toy, standing character, soft smooth lighting, soft pastel colors, skottie young, 3d blender render, polycount, modular constructivism, pop surrealism, physically based rendering, square image, centered -S1214204699 -C13.0 -Addim -s35

If you have any questions or comments, please feel free to reach out to me on Mastodon.

Jeremy


Post update

I was curious what MidJourney would do with that prompt, here are the 4 variations [needless to say MidJourney is still awesome!]:

Mastodon