{
  "job": "upscale",
  "want": "tile",
  "rank": 2,
  "short": "held by Tile, own model",
  "when": "enlarge with a lock: Tile holds the picture under a strong redraw, for much more detail; its own model copy can be a second model of the same family.",
  "title": {
    "zh": "Tile 锁图放大",
    "en": "Upscale held by Tile"
  },
  "about": {
    "zh": "放大时用 Tile 钉住画面，重画得更狠、细节更多。放大这一段有自己的一份模型（默认和底图同一个，LoRA 一起带上），想换第二个模型只要换这一格；Anima 上用 Anima 的 Tile LLLite。",
    "en": "A Tile holds the picture while it is enlarged and redrawn, so the redraw can go stronger: more detail. The redraw has its own copy of the model (the base's, its LoRAs too); a second model is one swap of that loader. On Anima, Anima's Tile LLLite holds it."
  },
  "source": {
    "zh": "参考工作流的「通用放大」组（平铺控制网 0.6，到 80% 的步数）和它的放大加载器；Anima 是你验证过的两段放大卡（潜空间 1.22 倍 + Tile LLLite，降噪 0.85）",
    "en": "The reference workflow's general upscale (tile ControlNet 0.6, to 80% of the steps) and its upscale loader; on Anima the user's tested two-pass card (latent 1.22x, Tile LLLite, denoise 0.85)"
  },
  "packs": {},
  "upscaler": true,
  "upscale_by": 1.25,
  "control": "tile",
  "extras": {
    "ColorMatch": "ComfyUI-KJNodes"
  },
  "nodes": {
    "own": {
      "class_type": "CheckpointLoaderSimple",
      "copy_of": "$sampler.model",
      "_meta": {
        "title": "Upscale model"
      },
      "inputs": {}
    },
    "upscaler": {
      "class_type": "UpscaleModelLoader",
      "inputs": {
        "model_name": "$upscaler"
      }
    },
    "enlarge": {
      "class_type": "ImageUpscaleWithModel",
      "inputs": {
        "upscale_model": [
          "upscaler",
          0
        ],
        "image": [
          "$save.images",
          "*"
        ]
      }
    },
    "fit": {
      "class_type": "ImageScaleBy",
      "inputs": {
        "image": [
          "enlarge",
          0
        ],
        "upscale_method": "lanczos",
        "scale_by": "$fit"
      }
    },
    "controlnet": {
      "class_type": "ControlNetLoader",
      "inputs": {
        "control_net_name": "$controlnet"
      }
    },
    "hold": {
      "class_type": "ControlNetApplyAdvanced",
      "inputs": {
        "positive": [
          "$sampler.positive",
          "*"
        ],
        "negative": [
          "$sampler.negative",
          "*"
        ],
        "control_net": [
          "controlnet",
          0
        ],
        "image": [
          "fit",
          0
        ],
        "vae": [
          "$decode.vae",
          "*"
        ],
        "strength": 0.6,
        "start_percent": 0.0,
        "end_percent": 0.8
      }
    },
    "encode": {
      "class_type": "VAEEncode",
      "inputs": {
        "pixels": [
          "fit",
          0
        ],
        "vae": [
          "$decode.vae",
          "*"
        ]
      }
    },
    "redraw": {
      "class_type": "KSampler",
      "like": "$sampler",
      "inputs": {
        "model": [
          "own",
          0
        ],
        "positive": [
          "hold",
          0
        ],
        "negative": [
          "hold",
          1
        ],
        "latent_image": [
          "encode",
          0
        ],
        "denoise": 0.5
      }
    },
    "decode": {
      "class_type": "VAEDecode",
      "inputs": {
        "samples": [
          "redraw",
          0
        ],
        "vae": [
          "$decode.vae",
          "*"
        ]
      }
    },
    "match": {
      "class_type": "ColorMatch",
      "optional": true,
      "inputs": {
        "image_ref": [
          "fit",
          0
        ],
        "image_target": [
          "decode",
          0
        ],
        "method": "mkl",
        "strength": 1.0,
        "multithread": true
      }
    },
    "save": {
      "class_type": "SaveImage",
      "inputs": {
        "images": [
          "match|decode",
          0
        ],
        "filename_prefix": "upscaled"
      }
    }
  },
  "variants": {
    "anima": {
      "packs": {
        "AnimaLLLiteApply_sdscripts": "ComfyUI-Anima-LLLite"
      },
      "upscaler": false,
      "control": null,
      "lllite": "tile",
      "nodes": {
        "own": {
          "class_type": "UNETLoader",
          "copy_of": "$sampler.model",
          "_meta": {
            "title": "Upscale model"
          },
          "inputs": {}
        },
        "encode": {
          "class_type": "VAEEncodeTiled",
          "inputs": {
            "pixels": [
              "$save.images",
              "*"
            ],
            "vae": [
              "$decode.vae",
              "*"
            ],
            "tile_size": 512,
            "overlap": 64,
            "temporal_size": 64,
            "temporal_overlap": 8
          }
        },
        "enlarge": {
          "class_type": "LatentUpscaleBy",
          "inputs": {
            "samples": [
              "encode",
              0
            ],
            "upscale_method": "bislerp",
            "scale_by": 1.22
          }
        },
        "tile": {
          "class_type": "AnimaLLLiteApply_sdscripts",
          "inputs": {
            "model": [
              "own",
              0
            ],
            "lllite_name": "$lllite",
            "image": [
              "$save.images",
              "*"
            ],
            "strength": 0.7,
            "start_percent": 0.0,
            "end_percent": 0.8,
            "preserve_wrapper": true
          }
        },
        "redraw": {
          "class_type": "KSampler",
          "like": "$sampler",
          "inputs": {
            "model": [
              "tile",
              0
            ],
            "positive": [
              "$sampler.positive",
              "*"
            ],
            "negative": [
              "$sampler.negative",
              "*"
            ],
            "latent_image": [
              "enlarge",
              0
            ],
            "steps": 30,
            "cfg": 4.4,
            "sampler_name": "dpmpp_2m_sde_gpu",
            "scheduler": "sgm_uniform",
            "denoise": 0.85
          }
        },
        "decode": {
          "class_type": "VAEDecodeTiled",
          "inputs": {
            "samples": [
              "redraw",
              0
            ],
            "vae": [
              "$decode.vae",
              "*"
            ],
            "tile_size": 512,
            "overlap": 64,
            "temporal_size": 64,
            "temporal_overlap": 8
          }
        },
        "match": {
          "class_type": "ColorMatch",
          "optional": true,
          "inputs": {
            "image_ref": [
              "$save.images",
              "*"
            ],
            "image_target": [
              "decode",
              0
            ],
            "method": "mkl",
            "strength": 1.0,
            "multithread": true
          }
        },
        "save": {
          "class_type": "SaveImage",
          "inputs": {
            "images": [
              "match|decode",
              0
            ],
            "filename_prefix": "upscaled"
          }
        }
      },
      "after": "On this Anima model the picture is enlarged 1.22x in latent space and redrawn at denoise 0.85, held by Anima's Tile LLLite (0.7, to 80% of the steps), with its own copy of the canvas's model and LoRAs (the loader titled Upscale model). For a second model (steadier, other skin or texture), set_model that loader to another Anima model; the LoRAs stay. Then run_workflow; the larger one is saved beside it (prefix upscaled). Compare face, hands and clothes with the base: Tile guides, it does not guarantee."
    }
  },
  "after": "run_workflow saves the picture as before and the 1.25x one beside it (prefix upscaled): a tile ControlNet holds it (strength 0.6, to 80% of the steps) while it is redrawn at denoise 0.5 with its own copy of the canvas's model and LoRAs (the loader titled Upscale model). For a second model (steadier, another look), set_model that loader to one of the same family (it reuses the first model's prompts); the LoRAs stay. Lower the denoise if it changes too much, raise the strength to hold it closer.",
  "gives": {
    "sampler": "redraw",
    "save": "save"
  },
  "own_model": {
    "nodes": [
      "own_1",
      "own"
    ],
    "copies_loras": true
  },
  "notes": "The source's tile values; denoise 0.5 because the tile holds the picture. The redraw's own model line replaced the separate second-model way (2026-11-04): same model by default, one swap for another. On Anima the user's tested route: Harem 1.5 base, Realskin 1.3 for the Tile upscale, its style LoRA on both, 832x1216 to 1016x1480, 116 s for the upscale on 6 GB.",
  "models": [
    "union_promax_sdxl",
    "noob_tile",
    "anima_tile_repair",
    "2x_animesharp_v4",
    "4x_ultrasharp"
  ],
  "tested": {
    "gpu": "RTX 4050 Laptop",
    "vram_gb": 6,
    "date": "2026-11-02",
    "runs": [
      {
        "model": "SDXL (Illustrious)",
        "size": "832×1216 → 1.25×",
        "seconds": 77
      },
      {
        "model": "Anima (Kirazuri → Realskin 1.3, Tile)",
        "size": "1216×832 → 1.22×",
        "seconds": 127
      }
    ]
  }
}