A3D | 3D Render Style LoRA

Update: Not yet a checkpoint, but close, with a dataset of 1954 pictures. You can create characters like Mercy, Tracer, Ahri without any additional LoRA. I have tested around 10 Character LoRAs using this style, and they have worked without any issues. However, since that's a small sample size, I would appreciate your help in testing more characters

There are no specific trigger words, plug-and-play.

The only factor that significantly affects the outcome is the strength of the LoRA. Therefore, I recommend experimenting with different LoRA strengths.

The style is quite powerful, resulting in every checkpoint containing the 3D effect. Its essentially a mini checkpoint.

If you wish to combine it with anime, for instance, you can decrease the strength to achieve the desired mix

models used

anime: YesMix - v2.0, MeinaAlter - V2

2,5D: DarkSun - v4.0, Animesh - Animesh-Pruned V2. ReV Animated - v1.2.2, PastelDiffusedMix - V2.1, MeinaUnreal - V3

realistic: Paragon V1.0

negative embeddings: bad_prompt, negative_hand, 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) from time to time im also using DPM++ 2M alt Karras

• Steps: 20 to 50

• 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

• Denoise strength upscaling: 0.3 to 0.55

If you want to autocomplete the tags & trigger im using, I recommend using Booru style tag autocompletion for AUTOMATIC1111's Stable Diffusion web UI

For full body shots i recommend using: ComfyUI Impact Pack

or the a1111 alternative GitHub - Bing-su/adetailer