Type | |
Stats | 1,409 |
Reviews | (212) |
Published | Jun 7, 2023 |
Base Model | |
Training | Steps: 3,072 Epochs: 16 |
Trigger Words | winterblessed diana |
Hash | AutoV2 7804EDC6B9 |
LoRA trained on Riven from League of Legends
requested by @bty_is_within_us
models used
anime: YabaL_Mix V3
2,5D: PastelDiffusedMix - V2.1, CamelliaMIx_2.5D - V2
realistic: Paragon V1.0
tags: winterblessed diana, 1girl,
optional tags: dress, pauldrons, long hair, white hair, forehead mark
negative embeddings: bad-hands-5, ng_deepnegative_v1_75t, bad-artist, EasyNegative
Recommended settings:
Make sure you are aware on the usage instructions of LORA
VAE: vae-ft-mse-840000-ema-pruned (for realistic models & 2,5D), kl-f8-anime2, blessed2.vae (for anime models)
Sampler: DPM++ SDE Karras or DPM++ 2M Karras(euler a for img2img upscaling)
Steps: 20 to 35
Clipskip: 2
Upscaler : GitHub - pkuliyi2015/multidiffusion-upscaler-for-automatic1111 4x-UltraSharp(for 2,5D & realistic), 4xAnimesharp(for anime)
CFG Scale : 5 to 9 (mostly 7)
LORA weight for txt2img: 0.5-1 (for the examples i mostly used 0.8)
Denoise strength upscaling: 0.4 to 0.7
for full body shots i recommend using: ComfyUI Impact Pack
or the a1111 alternative GitHub - Bing-su/adetailer