Raw²diance V2 Final / Turbo / V1
Turbo version of Raw²diance-V2
See Version Box.
Raw²diance V2 Final
Raw²diance V2 is an enhancement to V1 in every way. Its truly raw and sometimes rough but delivers pretty images. I wanted it to stand out from the crowd of merges and yes, it does it well.
Changes to V1:
Overall fine-tunig. Enhanced landscapes and lights. Better Cars and Bikes. Enhanced liveliness and artistic skills. Its still raw by design.
Start with: DMD2 4-Step Lora loaded
Sampler: LCM / CFG=1, Steps=10-16 (10 is sharpest - 16 is smooth & precise)
Non DMD2 Version since you need to load the DMD2 Lora anyway to tune when using more Loras and it can be used for Lora training. I still recommend DMD2 / LCM / CFG1 / 12 Steps. Use common settings for generation with other samplers. When using DMD2 Lora with other Loras together start at 0.85 for DMD2 and 0.65 for others.
Rawdiance V1 Final
Outputs crispy detailed and beautiful images while trying to look as less synthetic as possible. This was my goal from start. Yeah i know - synthetic robots in the samples. Pfff..
May post a DMD2 version if demanded.
Have fun and do not forget to post your creations and feedback.
Tip: You can load the DMD2 Lora with slightly negative weight for fine tuning output. (Necessary for stronger samplers like DPM). Lowering DMD2 weight needs higher CFG.
When using Character-Loras start with weights about 0.7.
Note regarding my prompts: I use Invoke, + is used to up weights like (word) = word+ . Stuff in <> are embeddings.
Feedback, suggestion and/or requests welcome.
WIP; Will add more samples over time...
Mirror and other Stuff: https://huggingface.co/nphSi/
Description
Turbo version of Raw²diance-V2.
Advantages: Since its tuned for Euler A it can use CFG and load Loras at weight up to around 1.0. Negative Prompting is supported.
Disadvantage: Less flexibility when using more aggressive schedulers like DPM.
Recommended settings:
Scheduler: Euler A, CFG: 1.3, Steps: 10
or
Scheduler: LCM, CFG: 1.8, Steps: 12
When using Euler A, Loras can be load with weight up to around 1.0. For LCM you may have to reduce it to 0.7 and also lower CFG. If you insist to use more aggressive schedulers like DPM/Karras you need to load DMD2-Lora with negative weights. I do not officially recommend/support this.