Verified: a year ago
SafeTensor
Type | |
Stats | 549 0 |
Reviews | (58) |
Published | Feb 21, 2024 |
Base Model | |
Hash | AutoV2 B3E31C7780 |
This Stability AI Model is licensed under the Stability AI Non-Commercial Research Community License, Copyright (c) Stability AI Ltd. All Rights Reserved.
Update v2.1: I've managed to remove the bias toward adding characters in general and superheroes in particular, making this a solid general digital painting model now.
Instructions for use:
This is a turbo model. I'm using DPM++ SDE at 8 steps with CFG set to 2. For hires fix and adetailer, I'm using 3-4 steps, although if you don't need super high resolution, it'll give you great quality on the first pass.
Test any quality-enhancing terms or LoRAs you put into your prompts and negative prompts. Don't just add "high quality" and "masterpiece" and expect it to be better without comparing the results on the same prompt and seed. I like the look of putting "photograph" in the negative prompt, but it works fine with it completely empty.