This is an Embedding built for Stable Diffusion 2.0.
Trained on 14 screenshots from the Sega Megadrive/Genesis game Sonic The Hedgehog.
Training was done with the Automatic1111 WebUI
Batch Size 7
Gradient Accumulation Steps 2
200 Steps (I will include previous step versions for you to try as well)
Phrase to invoke the Embedding is "MDSth". This was to avoid any interaction with other words and is shortened from MegaDrive Sonic The Hedgehog.
Default Generations with the "MDSth" phrase generate items that look like Sonic/MegaDrive era side scrollers:
Adding in other prompts can get interesting effects (Prompt: An Illustration of a Nissan GT-R by MDSth):
Prompt: A close-up Portrait Illustration of a woman in a red dress by MDSth
Also works well with Img2Img:
An example of the different step results (Results are dependent on prompt, your mileage may vary):