Generated with: Checkpoint: Js2Prony Lora: [SDXL] Artist Style: Takeda Hiromitsu/武田弘光/真珠貝 [for Js2Prony] by silvermoong (didn't use in this gallery but a VAE will produce better results)VAE: VAE for Pony Diffusion V6 XL
For people curious on how:
1. Get at least 4 GB VRAM 16 GB RAM, 17 GB hard disk space.
2. Install AUTOMATIC1111: Can google search for Installation on Windows 10/11 with NVidia-GPUs using release package
3. Download the checkpoint and lora and place them into the corresponding AUTOMATIC1111 folders checkpoints go into sd.webuiwebuimodelsStable-diffusion loras/lycorises go into sd.webuiwebuimodelsLora VAE goes into sd.webuiwebuimodelsVAE
One part that's not obvious is to use a Lora/lycoris, you go into the AUTOMATIC1111 Lora tab, click it and it will be added to the prompt with <loraname>
AUTOMATIC1111 needs to use that VAE in settings or if VAE set to automatic the VAE filename must be same as the checkpoint filename.
4. Run AUTOMATIC1111 and you can copy and paste the prompts, negative prompts, and settings from the civitai [SDXL] Artist Style: Takeda Hiromitsu/武田弘光/真珠貝 [for Js2Prony] by silvermoong gallery images
5. On an i7 processor with 4 GB VRAM 16 GB RAM, 20-28 steps(more steps for better quality), each image can be generated in 6 minutes or less with the correct settings. (it's recommended to have a NVIDIA video card with a lot of VRAM) If you only have 2-4 GB VRAM, open sd.webuiwebuiwebui-user.bat and change set COMMANDLINE_ARGS= to set COMMANDLINE_ARGS= --xformers --lowvram (-xxformers is optimizations for NVIDIA cards and the --lowvram optimizes for processing with low VRAM)