Type | |
Stats | 641 |
Reviews | (81) |
Published | Jun 1, 2023 |
Base Model | |
Training | Steps: 2,496 Epochs: 16 |
Trigger Words | welt yang |
Hash | AutoV2 774DC2A731 |
LoRA trained on Welt Yang from Honkai: Star Rail
requested by @TwelveSH
models used
anime: YabaL_Mix V3(image 2), MeinaAlter(image 6), Anything and Everything(image 1,3,4,5)
2,5D: HELLmix(image 7,8), DarkSun, MeinaUnreal
tags: welt yang, 1boy,
optional tags: male focus, glasses, grey coat, brown scarf, white shirt, multicolored hair, brown hair,
negative embeddings: embedding:bad-hands-5, embedding:ng_deepnegative_v1_75t, embedding:bad-artist, embedding: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
Steps: 20 to 35
Clipskip: 2
Upscaler : img2img in comfyui using 4x-UltraMix_Smooth
CFG Scale : 5 to 9 (mostly 7 and 8)
LORA weight for txt2img: 0.5-1 (for the examples i mostly used 0.8)
Denoise strength for img2img: 0.4 to 0.7
for the pictures i used comfyui