Model

Anime4KCPP V3 supports the following CNN model variants:

• acnet-gan: Lightweight CNN, detail enhancement.
• acnet-hdn0: Lightweight CNN, mild denoising.
• acnet-hdn1: Lightweight CNN, moderate denoising.
• acnet-hdn2: Lightweight CNN, heavy denoising.
• acnet-hdn3: Lightweight CNN, extreme denoising.
• arnet-hdn: Lightweight ResNet, mild denoising.