Anima 三次采样 + ControlNet-LLLite + 4K 高清放大重绘工作流

Overview 概述

English: A complete production pipeline for the Anima base model, built to solve Anima's two core weaknesses — soft detail and structural drift at high resolution. The workflow fuses three Anima checkpoints, applies two LoRA stacks, runs three sequential sampling passes with ControlNet-LLLite structural control, then performs a 4K Ultimate SD Upscale with a dedicated upscale model, finishing with RTX denoise/deblur and anime color grading. Designed for clean, high-detail, structurally-stable anime output.

中文： 一套针对 Anima 底模的完整生产管线，专为解决 Anima 两大核心短板——细节偏软、高分辨率下结构漂移。工作流融合三个 Anima 底模、叠加两组 LoRA 栈、串联三次采样并配合 ControlNet-LLLite 结构控制，最后进行4K Ultimate SD Upscale 超分（专用放大模型），以 RTX 降噪/去模糊与动漫调色收尾。产出干净、高细节、结构稳定的动漫成品。

Architecture 架构（三段式采样 + 4K 超分）

English:

• Stage 1 — Base Generation（一采）: 40 steps, CFG 5, denoise 1.0, at 864×1536 (portrait) or 1536×864 (landscape).

• Stage 2 — Latent Upscale & Redraw（二采）: bicubic latent upscale ×1.5, then 40 steps at denoise 0.8.

• Stage 3 — ControlNet-LLLite Guided Redraw（三采）: 40 steps at denoise 0.8, guided by depth/lineart LLLite to lock structure and sharpen detail.

• Stage 4 — 4K Ultimate SD Upscale（4K超分）: 2× upscale with 4x-AnimeSharp model, 30 steps at denoise 0.05, tile 2304×1296 (full-image mode) for a clean high-res finish.

• Post-processing: RTX Denoise → RTX Deblur → EasyColorCorrection (Anime Moody).

中文：

• 第一采——基础生成： 40 步、CFG 5、denoise 1.0，在 864×1536（竖版）或 1536×864（横版）生成初始图。

• 第二采——潜空间放大重绘： bicubic 潜空间放大 1.5 倍，再以 denoise 0.8 采样 40 步。

• 第三采——ControlNet-LLLite 引导重绘： denoise 0.8 采样 40 步，深度/线稿 LLLite 控制锁定结构、锐化细节。

• 第四段——4K Ultimate SD Upscale： 4x-AnimeSharp 放大模型 2 倍超分，30 步、denoise 0.05、tile 2304×1296（整图模式），产出干净高分辨率成片。

• 后处理： RTX 降噪 → RTX 去模糊 → EasyColorCorrection（Anime Moody 预设）。

Key Features 核心特性

English:

1. Triple-checkpoint fusion — chosenMixAnima_v10 + anima_aestheticV11 + miaomiaoHarem_anima16, merged via ModelMergeSimple (0.6) + ModelMergeBlocks (input 0.9 / middle 0.2 / out 0.7) for layered control of structure, style and detail.

2. Dual LoRA stacks — quality stack (masterpiece 0.9 + highres-aesthetic-boost 0.9 + rdbt 0.9 + light-concepts 0.6) + detail stack (rendering_detailer −0.3 + context_detailer 0.3).

3. ControlNet-LLLite structural control — depth + lineart LLLite keep anatomy stable across all sampling passes.

4. ReferenceLatent anchoring — preserves the generated image as a structural anchor during redraw.

5. AuraFlow shift=3 — compensates Anima's native softness for sharper lines and richer texture.

6. 4K Ultimate SD Upscale — dedicated 4x-AnimeSharp model + full-image tile mode for clean 2× upscale.

7. Camera control node — azimuth/elevation/distance-driven camera prompt injection.

8. RTX + color grading post-chain — cleans artifacts and applies an Anime Moody grade.

中文：

1. 三底模融合——chosenMixAnima_v10 + anima_aestheticV11 + miaomiaoHarem_anima16，经 ModelMergeSimple(0.6) + ModelMergeBlocks(input 0.9/middle 0.2/out 0.7) 分层控制结构、风格与细节。

2. 双 LoRA 栈——质量栈（masterpiece 0.9 + highres-aesthetic-boost 0.9 + rdbt 0.9 + light-concepts 0.6）+ 细节栈（rendering_detailer −0.3 + context_detailer 0.3）。

3. ControlNet-LLLite 结构控制——深度 + 线稿 LLLite 全程锁定解剖结构。

4. ReferenceLatent 锚定——重绘时以已生成图为结构锚，防止漂移。

5. AuraFlow shift=3——补偿 Anima 原生偏软，线条更锐、纹理更实。

6. 4K Ultimate SD Upscale——专用 4x-AnimeSharp 模型 + 整图 tile 模式，干净 2 倍超分。

7. 相机控制节点——方位角/俯仰角/距离驱动的相机提示词注入。

8. RTX + 调色后处理链——清理伪影并套用 Anime Moody 动漫调色。

Model & LoRA Dependencies 依赖清单

Type 类型File 文件Weight 权重CheckpointchosenMixAnima_v10.safetensors(merged)Checkpointanima_aestheticV11.safetensors(merged)CheckpointmiaomiaoHarem_anima16.safetensors(merged)LoRAanima-base-1-masterpiece-v510.9LoRAanima-highres-aesthetic-boost0.9LoRArdbt_v2.1_base_anima_b10.9LoRAanima-base-1-light-concepts-v310.6LoRArendering_detailer_base10−0.3LoRAanima_context_detailer_base100.3CLIPmiaomiaoHarem_anima16_txt.safetensors—VAEqwen_image_vae.safetensors—Upscale Model4x-AnimeSharp.safetensors—LLLiteanima-lllite-depth-10.8LLLitecontrolnetLlliteAnima_lineart0.8

Settings 参数配置

English:

• Sampler: euler_ancestral / Scheduler: normal

• Steps: 40 (UltimateSDUpscale: 30) / CFG: 5

• Denoise: Stage1 1.0 → Stage2 0.8 → Stage3 0.8 → 4K pass 0.05

• Resolution: 864×1536 portrait / 1536×864 landscape

• Latent upscale: bicubic ×1.5

• UltimateSDUpscale: 2× / tile 2304×1296 / padding 128

• ModelMergeSimple: 0.6 / ModelMergeBlocks: input 0.9, middle 0.2, out 0.7

• ModelSamplingAuraFlow: shift=3

• Seed: fixed, with increment mode available

中文：

• 采样器：euler_ancestral / 调度器：normal

• 步数：40（UltimateSDUpscale 用 30）/ CFG：5

• Denoise：一采 1.0 → 二采 0.8 → 三采 0.8 → 4K 超分 0.05

• 分辨率：864×1536 竖版 / 1536×864 横版

• 潜空间放大：bicubic ×1.5

• UltimateSDUpscale：2× / tile 2304×1296 / padding 128

• ModelMergeSimple：0.6 / ModelMergeBlocks：input 0.9, middle 0.2, out 0.7

• ModelSamplingAuraFlow：shift=3

• 种子：固定，支持递增模式

How to Use 使用步骤

English:

1. Install required custom node packs (rgthree, KJ nodes, LLLite sd-scripts, RTX nodes, EasyColorCorrection, Layer Utility, Ultimate SD Upscale).

2. Download all checkpoints/LoRA/LLLite/upscale model files to correct folders.

3. Load the workflow and fix missing model paths.

4. Write your prompt (English tags + natural language works best with Anima's Qwen encoder).

5. Run. Adjust LoRA weights, merge ratios and denoise as needed.

中文：

1. 安装所需自定义节点包（rgthree、KJ nodes、LLLite sd-scripts、RTX 节点、EasyColorCorrection、Layer Utility、Ultimate SD Upscale）。

2. 下载全部底模/LoRA/LLLite/放大模型文件到对应文件夹。

3. 加载工作流并修正缺失路径。

4. 编写提示词（Anima 的 Qwen 编码器最适合英文标签 + 自然语言）。

5. 运行，按需调整 LoRA 权重、融合比例和 denoise。