{
  "job": "detail",
  "want": "hands",
  "rank": 2,
  "short": "enlarge, then hands (fallback)",
  "enlarge": {
    "megapixels": 4.2,
    "most": 2.0,
    "unknown": 1.5
  },
  "title": {
    "zh": "修手·先放大再修（备选）",
    "en": "Fix hands: enlarge, then fix (fallback)"
  },
  "about": {
    "zh": "整张图放大到约 400 万像素（小图 2 倍）轻轻重画，让手变大；再找出每只手，放大重画后贴回去。",
    "en": "The whole picture is enlarged to about 4 megapixels (2× for a small one) and redrawn lightly, so the hands grow; then each hand is found, redrawn larger and pasted back."
  },
  "source": {
    "zh": "参考作者的「先放大重画」和「手部精修」两步合在一起，2026-10-29 在 RTX 4050 上试出这组数",
    "en": "The reference author's light redraw after enlarging and his hand detailing, put together; these numbers tried on an RTX 4050 on 2026-10-29"
  },
  "packs": {
    "FaceDetailer": "ComfyUI-Impact-Pack",
    "UltralyticsDetectorProvider": "ComfyUI-Impact-Subpack"
  },
  "takes": "picture",
  "detectors": {
    "hand": "hand_yolov8"
  },
  "roles": {
    "sampler": {
      "denoise": 0.4
    }
  },
  "nodes": {
    "picture": {
      "class_type": "LoadImage",
      "inputs": {
        "image": "$image"
      }
    },
    "bigger": {
      "class_type": "ImageScaleBy",
      "inputs": {
        "image": [
          "picture",
          0
        ],
        "upscale_method": "lanczos",
        "scale_by": "$enlarge"
      }
    },
    "encode": {
      "class_type": "VAEEncode",
      "inputs": {
        "pixels": [
          "bigger",
          0
        ],
        "vae": [
          "$decode.vae",
          "*"
        ]
      },
      "wire_to": "$sampler"
    },
    "hand_finder": {
      "class_type": "UltralyticsDetectorProvider",
      "inputs": {
        "model_name": "$hand"
      }
    },
    "hands": {
      "class_type": "FaceDetailer",
      "like": "$sampler",
      "inputs": {
        "image": [
          "$decode",
          0
        ],
        "model": [
          "$sampler.model",
          "*"
        ],
        "clip": [
          "$sampler.positive.clip",
          "*"
        ],
        "vae": [
          "$decode.vae",
          "*"
        ],
        "positive": [
          "$sampler.positive",
          "*"
        ],
        "negative": [
          "$sampler.negative",
          "*"
        ],
        "bbox_detector": [
          "hand_finder",
          0
        ],
        "guide_size": 768,
        "max_size": 1024,
        "denoise": 0.45,
        "bbox_threshold": 0.3,
        "bbox_crop_factor": 2.0,
        "feather": 5,
        "drop_size": 10,
        "wildcard": "detailed hand"
      }
    },
    "save": {
      "class_type": "SaveImage",
      "inputs": {
        "images": [
          "hands",
          "image"
        ],
        "filename_prefix": "hands_fixed"
      }
    }
  },
  "after": "A fallback: a FLUX-Fill redraw of one square crop around each hand did better in the user's tests; use that when you have it. The picture is enlarged to about 4 megapixels (2× for one of about 1 megapixel, at most 2×, not at all for a large one) and drawn again lightly (the sampler now takes it, at denoise 0.4), then every hand found is redrawn larger (hands, denoise 0.45) and saved with prefix hands_fixed; the enlarged one is saved too, without the hand pass. About 3 tries in 4 give a good hand: run_workflow with batch 4 (each try its own seed), then check them before the user does: zoom_image each hands_fixed picture on the hand, compare_with the old picture, and drop a hand with the wrong number of fingers, fused or melted fingers, or one far bigger or smaller than its arm. None good: run_workflow batch 4 again, at most twice. Then a picks panel zoomed on the hand lets the user choose among the good ones.",
  "notes": "Tried on the Codex test picture (1280×832, three characters, the hand about 110×78 px): the hand pass alone gave a good hand in about half the seeds whatever its size (512 or 768) and denoise (0.45 or 0.55); enlarging 1.5× first gave good hands in 3 of 4, 2× gave sharper fingers in the same seeds (about 105 s instead of 75 s). Denoise 0.4 for the enlarged redraw kept the layout; faces and splashes change a little.",
  "models": [
    "hand_yolov8s",
    "anima_turbo_lora"
  ],
  "tested": {
    "gpu": "RTX 4050 Laptop",
    "vram_gb": 6,
    "date": "2026-10-29",
    "runs": [
      {
        "model": "Anima (Kirazuri v4, Turbo LoRA)",
        "size": "1280×832 → 2560×1664",
        "seconds": 105
      }
    ]
  },
  "variants": {
    "anima": {
      "loras": {
        "turbo": "anima-turbo-lora"
      },
      "without_loras": "On an Anima model the Anima Turbo LoRA (anima-turbo-lora-v0.2, the user can get it in the ways panel) makes this about three times faster: 10 steps at CFG 1.",
      "roles": {
        "sampler": {
          "denoise": 0.4,
          "steps": 10,
          "cfg": 1.0,
          "sampler_name": "euler_ancestral",
          "scheduler": "normal"
        }
      },
      "nodes": {
        "picture": {
          "class_type": "LoadImage",
          "inputs": {
            "image": "$image"
          }
        },
        "turbo": {
          "class_type": "LoraLoaderModelOnly",
          "inputs": {
            "model": [
              "$sampler.model",
              "*"
            ],
            "lora_name": "$turbo",
            "strength_model": 0.7
          },
          "wire_to": "$sampler"
        },
        "bigger": {
          "class_type": "ImageScaleBy",
          "inputs": {
            "image": [
              "picture",
              0
            ],
            "upscale_method": "lanczos",
            "scale_by": "$enlarge"
          }
        },
        "encode": {
          "class_type": "VAEEncode",
          "inputs": {
            "pixels": [
              "bigger",
              0
            ],
            "vae": [
              "$decode.vae",
              "*"
            ]
          },
          "wire_to": "$sampler"
        },
        "hand_finder": {
          "class_type": "UltralyticsDetectorProvider",
          "inputs": {
            "model_name": "$hand"
          }
        },
        "hands": {
          "class_type": "FaceDetailer",
          "like": "$sampler",
          "inputs": {
            "image": [
              "$decode",
              0
            ],
            "model": [
              "turbo",
              0
            ],
            "clip": [
              "$sampler.positive.clip",
              "*"
            ],
            "vae": [
              "$decode.vae",
              "*"
            ],
            "positive": [
              "$sampler.positive",
              "*"
            ],
            "negative": [
              "$sampler.negative",
              "*"
            ],
            "bbox_detector": [
              "hand_finder",
              0
            ],
            "guide_size": 768,
            "max_size": 1024,
            "steps": 10,
            "cfg": 1.0,
            "sampler_name": "euler_ancestral",
            "scheduler": "normal",
            "denoise": 0.45,
            "bbox_threshold": 0.3,
            "bbox_crop_factor": 2.0,
            "feather": 5,
            "drop_size": 10,
            "wildcard": "detailed hand"
          }
        },
        "save": {
          "class_type": "SaveImage",
          "inputs": {
            "images": [
              "hands",
              "image"
            ],
            "filename_prefix": "hands_fixed"
          }
        }
      },
      "after": "A fallback: a FLUX-Fill redraw of one square crop around each hand did better in the user's tests; use that when you have it. On this Anima model the Anima Turbo LoRA (0.7) makes both passes 10 steps at CFG 1, about 105 s for 1280×832 (2×) on a 6 GB card. The picture is enlarged to about 4 megapixels (2× for one of about 1 megapixel, at most 2×, not at all for a large one) and drawn again lightly (the sampler now takes it, at denoise 0.4), then every hand found is redrawn larger (hands, denoise 0.45) and saved with prefix hands_fixed; the enlarged one is saved too, without the hand pass. About 3 tries in 4 give a good hand: run_workflow with batch 4 (each try its own seed), then check them before the user does: zoom_image each hands_fixed picture on the hand, compare_with the old picture, and drop a hand with the wrong number of fingers, fused or melted fingers, or one far bigger or smaller than its arm. None good: run_workflow batch 4 again, at most twice. Then a picks panel zoomed on the hand lets the user choose among the good ones."
    }
  }
}