Anima settings: sampler eulera / Euler Ancestral, scheduler normal , denoise 0.65 .
Perfil de execução
Descrição da fonte
Anima settings: sampler euler_a / Euler Ancestral, scheduler normal , denoise 0.65 .
Known limitation: with Anima, this workflow currently works reliably only with square images for reasons not yet determined. Portrait inputs may produce blurry results.
Estimativa de requisito VRAM
Estimativa indisponível
242 MB distribuídos em 1 de 3 arquivos de modelo. Total de arquivos do modelo + 25% de overhead de carregamento + 2 GB de buffer de execução, arredondado para cima.
Requisitos
Requisitos 4Comentário gerado por IA
Explicação gerada por IA com base nos detalhes da fonte e da configuração. As sugestões são claramente identificadas.
Pixel Anchored Remaster is an Anima image-remaster workflow that takes a base image, upscales it 2×, and uses a lower-denoise detail pass to reduce upscale artifacts and produce a cleaner image while keeping the composition anchored.
The image is normalized to a fixed canvas, re-encoded into latent space, then passed through a second KSampler at lower denoise to rebuild detail.
The v1.2 notes connect HiRes Fix output directly to the final Pixel Anchored Remaster stage.
Start with a good original image; the described flow can run it through HiRes Fix before the final remaster stage.
Sugestão · não verificado
The intended output is a cleaner 2× remaster with controlled drift, not strict pixel preservation or a 4× upscale.
Install the PixelAnchoredRemaster custom node from https://github.com/NotProniss/PixelAnchoredRemaster or through the custom node manager.
The listed model files are anima_baseV10_txt.safetensors (text encoder), qwen_image_vae.safetensors (VAE), and waiANIMA_v10Base10.safetensors (checkpoint).
The workflow requires the PixelAnchoredRemaster node pack, listed as pixel-anchored-remaster.
Before use, check that these filenames match the text encoder, VAE, and checkpoint files in your setup.
Sugestão
The v1.3 Anima settings are sampler euler_a (Euler Ancestral), scheduler normal, and denoise 0.65.
The v1.2 notes specify bislerp for final latent upscaling and 0.55 for final remaster denoise.
With Anima, prefer square inputs; the notes say portrait inputs may produce blurry results.
Sugestão · não verificado
Isso precisa ser editado?
Faça login para enviar uma solicitação de edição.
Fontes
1 fonteTrechos de fonte
3 trechosSource context: 720 downloads · Type Workflows · Base model Anima
anima_baseV10_txt.safetensors
Não resolvidoText encoder · Unknown
VAE · 242 MB · Civitai · 2641326
waiANIMA_v10Base10.safetensors
Não resolvidoCheckpoint · Civitai · 2544636
Pacote de nós · Registry
Source context: 717 downloads · Type Workflows · Base model Anima
Source context: 533 downloads · Type Workflows · Base model Anima