Type | |
Stats | 170 0 |
Reviews | (5) |
Published | Feb 17, 2024 |
Base Model | |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 D5A34D93A9 |
This model and its variant were made via merging several datasets and LORAs of panoramas, landscapes and attractive women. It was originally for personal use and I never kept track of what was added or removed over time.
Variants
The two provided variants are tuned for dark/dramatic lighting (MMUltraPan_A) and a more neutral less dramatic variant (MMUltraPan_B). Contrast, tone and details are the only differences between the variants. If still too dark, prompting keywords "day, bright" and similar can brighten up the image as shown in some examples.
All example panoramic images posted were generated straight-shot with no LORAs/embeddings/HIRES/upscale/controlnet or any other changes whatsoever to the generated images.
All example images of women posted were generated with no LORAs/controlnet/adetailer and simply have negative embeddings and HIRESfix upscale (SwinIR with tile size 96). The prompts were created by LLM.