When you train a LoRA for stable diffusion, you can have it create intermediate models at certain points during the training. That way you can test each model to see at what point the model has a good fit. Otherwise you might end up with a model that is overfitted (i.e., can only produce images that are very similar to the training dataset, regardless of the prompt or that producegarbled images). I tested every variant of each model with the same prompts. Many of the images were rubbish, but there were a few that were worth keeping, even though I didn't ultimately keep the model that I used to generate them—often because they were overfitted.
Here follows are some additional images that I generated while testing the different intermediate models during the training of my LoRA for recreating the style of Midjourney v1. Some of them are the initial image produced, others are the upscaled version, and in a few cases I'll share both.
MadeJourney
Cthulhu
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
CloudJourney
Cthulhu
American Flag
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
GritJourney
Cthulhu
American Flag
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
SwathJourney
Cthulhu
American Flag
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
ScribJourney (first attempt)
Cthulhu
American Flag
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
ScribJourney (second attempt)
Cthulhu
Rocket
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
TwistJourney and BlobJourney
Cthulhu
American Flag
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
StarkJourney
Cthulhu
American Flag
Rocket
Floating Pyramid
Alien Landscape
Blue Monster
Celestial Library
Angry Man
Samurai
Goat
Burning City
*Journey (combinations)
Cthulhu
These illustrations were drawn using Stable Diffusion 1.5 as implemented in SD.Next, with the DreamShaper 7 model.
The final LoRAs can be downloaded at https://civitai.com/models/253022/madejourney.
