Base model: Anything v3.Please add vr-hoshimi tag to the prompts while using. No special tags exist.2D ver. gives more stable outputs, while 2D+3D ver. will be better for generating full body images of different poses.Latent upscaling could be used for better details on the face for mixed version.All right reserved to original character. Please contact me for deletion.

Files with SHA256: eb61408e30b5b55d44c5f8611169376c5fa398a8668988b0a8def0e76d610a02