Duplicate from Noob-Master60nine/Anima-Control-Pose

Co-authored-by: Adam Madrigal <Noob-Master60nine@users.noreply.huggingface.co>

.gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+samples/grid.png filter=lfs diff=lfs merge=lfs -text

LICENSE

@@ -0,0 +1,21 @@
+Anima Control — Pose (Preview-1)
+Non-Commercial License (inherited)
+
+These adapter weights (adapter_model.safetensors) are a derivative of the Anima
+base model (circlestone-labs/Anima), which is itself a derivative of
+nvidia/Cosmos-Predict2-2B-Text2Image.
+
+As a derivative, these weights inherit and are bound by the upstream terms:
+
+  1. CircleStone Labs Non-Commercial License
+     https://huggingface.co/circlestone-labs/Anima  (see its LICENSE.md)
+  2. NVIDIA Open Model License Agreement (insofar as it applies to Derivative Models)
+     https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/
+
+The MODEL WEIGHTS are for NON-COMMERCIAL use only.
+
+Generated images (Outputs) are NOT restricted by these terms and may be used
+commercially, consistent with the base model's license.
+
+This file is a pointer to the governing upstream terms, not an independent grant
+of rights. If an upstream license forbids a use, that prohibition applies here.

README.md

@@ -0,0 +1,187 @@
+---
+license: other
+license_name: circlestone-labs-non-commercial-license
+license_link: LICENSE
+base_model: circlestone-labs/Anima
+library_name: diffusers
+tags:
+  - control-lora
+  - pose
+  - dwpose
+  - anima
+  - cosmos-predict2
+  - comfyui
+  - text-to-image
+pipeline_tag: text-to-image
+---
+
+# Anima Control — Pose (Preview-2)
+
+> ⚠️ **Preview-2 — still experimental.** Better than Preview-1, but not finished. It will still
+> miss poses and produce deformed bodies, fused hands, and similar artifacts. Treat it as a
+> work-in-progress preview, not a production tool. **Non-commercial use only** (inherits the Anima
+> base model license). Behaviour and weights may still change.
+
+A native **pose control adapter** for the Anima v1.0 image model: condition generation on a
+skeleton pose map so the subject follows a target pose. Preview-2 is the same idea as Preview-1,
+trained at higher resolutions on a larger corpus, and shipped with a friendlier ComfyUI node.
+
+## What changed since Preview-1
+
+- **Multi-resolution: 512, 768 and 1024.** Preview-1 was 512-only; this runs at all three, and the
+  bodies hold together much better at the larger sizes. 1024 looks best.
+- **Larger training corpus** than Preview-1's ~3,900 examples, which shows up as cleaner anatomy and
+  steadier poses.
+- **Better pose-following at every resolution**, with the biggest gains at 768 and 1024.
+- **New ComfyUI node, "Anima Pose Control."** Drop in a reference photo; it detects the pose for you
+  and renders the skeleton, and you can pick how the skeleton is drawn (thin, thick, puppet, heatmap,
+  or with hands/face stripped). The plain thin skeleton is still the best default — it's the only one
+  the model trained on — but a different render occasionally lands a stubborn pose when the seed
+  alone won't.
+
+## Method
+
+Unchanged from Preview-1: a channel-concat control-LoRA on the frozen Anima DiT.
+
+**Conditioning.** The base VAE encodes the skeleton pose map into a control latent, the same latent
+space as the noisy image, so the control stays spatially aligned with the generation.
+
+**Fusion (`ControlEmbedder` + `ControlInitialLayer`).** A zero-initialized `ControlEmbedder` produces
+control tokens that are **added to the frozen base patch-embed output**. Zero-init means training
+starts as an exact no-op (output == base) and the control contribution grows only as it earns loss,
+so at `strength = 0` the adapter is exactly the base model.
+
+**Trainable parameters.** The `ControlEmbedder` plus a rank-16 low-rank adapter on the transformer
+blocks. The base transformer, text encoder, and VAE stay frozen.
+
+```
+skeleton ─▶ VAE ─▶ control latent ─┐
+                                   ▼  (+ zero-init ControlEmbedder)
+noisy latent ─▶ patch-embed ─▶ [ControlInitialLayer] ─▶ Block×N (+ rank-16 LoRA) ─▶ output
+```
+
+## Training
+
+**Data.** (image, skeleton, caption) triples generated by Anima from a broad prompt distribution;
+skeletons rendered from each image's detected keypoints (DWPose, COCO-WholeBody, black background).
+Preview-2 uses a substantially larger corpus than Preview-1.
+
+| Setting | Value |
+|---|---|
+| Resolution | **512 + 768 + 1024**, aspect-ratio bucketed |
+| Adapter rank | 16 |
+| Learning rate | 1e-4 |
+| Epochs | 8 |
+| Control dropout | 0.1 |
+| Precision | bf16 |
+
+> Final training loss ≈ 0.11 (denoising MSE, mean over the final 400 steps).
+
+## Results
+
+Measured on held-out full-body poses (fresh generations, not seen in training). Pose agreement is
+body-PCK@0.1: re-detect keypoints on each output, compare to the target skeleton.
+
+![Pose control across skeleton styles](samples/grid.png)
+
+*Each grid: the BASE column shows the reference and the no-control generation (same prompt, different
+seed — it ignores the pose); the remaining columns show the skeleton, drawn in each style, over the
+pose-controlled output. Control follows the pose; no-control doesn't.*
+
+| body-PCK@0.1 | control off | control on |
+|---|---|---|
+| 512 | ~0.33 | **~0.59** |
+| 768 | ~0.38 | **~0.67** |
+| 1024 | ~0.37 | **~0.83** |
+
+Preview-1 reached ~0.59 at 512. Preview-2 matches that at 512 and pulls clearly ahead at 768 and
+1024 — the gains grow with resolution.
+
+**What's still off (honest):**
+- It doesn't always follow the skeleton, even a clean, correct one.
+- A thin stick figure on black isn't how anime is drawn, so the model only half-reads it; the worst
+  artifacts (fused hands, mush) cluster where the skeleton is busiest.
+- Dynamic poses — running, jumping, sitting — are the least reliable.
+- On short or vague prompts the style flattens toward a samey default; a richer prompt fixes it.
+
+## Usage (ComfyUI)
+
+Preview-2 uses two small custom nodes: **Anima Control Apply** (`AnimaControlApply`) applies the
+adapter, and **Anima Pose Control** (`AnimaPoseControl`) detects the pose from a photo and renders
+the skeleton for you.
+
+### Install
+
+1. Download `anima_pose_preview2.safetensors` into `ComfyUI/models/loras/`.
+2. Copy both folders from `comfyui/` in this repo into `ComfyUI/custom_nodes/`:
+   `anima_control_lora/` and `ComfyUI-anima-pose-control/`. Restart ComfyUI. ComfyUI-Manager installs
+   the second node's `requirements.txt` automatically; otherwise:
+   `pip install -r ComfyUI-anima-pose-control/requirements.txt` (rtmlib, opencv-python, onnxruntime,
+   numpy; torch and Pillow come with ComfyUI).
+3. Load a workflow from the menu. `pose_control_demo.json` is the easiest — control vs no-control side
+   by side. Also included: `pose_control.json` (simple), `pose_control_edit.json` (single node, pick
+   the skeleton style), `pose_control_compare.json` (one pose across every style at once).
+
+`anima_pose_preview2.safetensors` holds both the low-rank adapter (`lora.*` keys) and the control
+embedder (`control_embedder.*` keys); `LoraLoaderModelOnly` reads the first set and Anima Control
+Apply reads the second, both pointing at the same file.
+
+**Strength.** `0.0` is the base model with no control; `1.0` follows the skeleton (range 0–2). Higher
+tracks the pose more closely but can cost some image quality.
+
+### Pose detector (first run)
+
+The Anima Pose Control node needs a pose detector (rtmlib). On first use it downloads two ONNX
+files (~316 MB) from this repo's `detector/` folder and caches them under
+`~/.cache/rtmlib/hub/checkpoints/`; after that it runs offline. If you see
+`urllib ... getaddrinfo failed`, your machine couldn't reach the download host — download
+`detector/yolox_m_8xb8-300e_humanart-c2c7a14a.onnx` and
+`detector/rtmw-dw-x-l_simcc-cocktail14_270e-256x192_20231122.onnx` from the Files tab by hand and
+drop both into that cache folder (create it if missing), then restart ComfyUI.
+
+## Roadmap
+
+**Preview-3** targets the headline weakness: the model only half-reads the skeleton. The thin
+lines-on-black signal seems foreign to it, so a representation bake-off is testing other renders
+(thicker, puppet/segmentation, depth-mannequin) to find what Anima follows best, plus the smaller
+fixes (default strength, dropping noisy hand points, more varied captions). The other half is the
+detector: the one used here was built for photos, not anime, so on art it produces noisy skeletons.
+Preview-3 will most likely wait on a purpose-built **anime pose detector** ("DWPose for anime"),
+then retrain on the winning representation with a lot more dynamic-pose data.
+
+**Version 1.0** comes after Preview-3, if it lands clean — the first non-preview release, trained on
+a good deal more data again.
+
+Pose is the first component on a shared control harness for Anima; planned siblings are image-prompt
+(IP-Adapter) and face-identity conditioning.
+
+## License
+
+These weights are a derivative of the Anima base model
+([`circlestone-labs/Anima`](https://huggingface.co/circlestone-labs/Anima)) and inherit its terms:
+the **CircleStone Labs Non-Commercial License**, and — because Anima is itself a derivative of
+Cosmos-Predict2 — the [NVIDIA Open Model License](https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/).
+
+The **model weights are for non-commercial use only.** Generated images (outputs) are not restricted
+by these terms and may be used commercially. See the bundled [`LICENSE`](./LICENSE) for the full text.
+
+## Support
+
+Building these models means mining and labeling a lot of images and renting GPUs to train on them. If
+they're useful to you and you want to chip in, it's appreciated and never expected:
+https://ko-fi.com/claquasse
+
+## Citation
+
+```bibtex
+@misc{anima_control_pose_preview2,
+  title  = {Anima Control --- Pose (Preview-2)},
+  author = {Claquasse},
+  year   = {2026},
+  note   = {Preview-2 multi-resolution pose control adapter for Anima v1.0},
+  howpublished = {\url{https://huggingface.co/Claquasse/Anima-Control-Pose}}
+}
+```
+
+Built on Anima (CircleStone Labs), the Cosmos-Predict2 transformer architecture, and the
+diffusion-pipe training framework.

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:976c93669dd69eb5b4d7eb63ad22c344b69860b23f8977623b70a6ba344027f4
+size 75051096

anima_control_tools.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcd28519425a96b85ab5a52bf99014519812c81fca6c863fc72d7391c5dabcb1
+size 28067

anima_pose_preview2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c10fa11628e9b1bf16dfe095ac5c6dc4ddd9ae77d7fe023d0a1857f7e29c516f
+size 75051096

comfyui/ComfyUI-anima-pose-control/README.md

@@ -0,0 +1,34 @@
+# ComfyUI-anima-pose-control
+
+A single ComfyUI node — **Anima Pose Control** (`AnimaPoseControl`) — that detects a pose from a photo,
+lets you drag the body joints on a mini-canvas in the node, picks a predefined skeleton style, paints the
+styled skeleton inline, and outputs the control image for a pose-control LoRA.
+
+## Install
+Clone into `ComfyUI/custom_nodes/` and restart ComfyUI (ComfyUI-Manager installs `requirements.txt` for you):
+```bash
+cd ComfyUI/custom_nodes
+git clone <this repo> ComfyUI-anima-pose-control
+ComfyUI/venv/bin/python -m pip install -r ComfyUI-anima-pose-control/requirements.txt
+```
+Dependencies: `rtmlib` (DWPose Wholebody-133 detection — same model/convention as the training data),
+`opencv-python`, `onnxruntime`, `numpy`. torch / Pillow are provided by ComfyUI.
+
+## Node: Anima Pose Control
+**Inputs:** `image` (optional) · widgets `style` (R0_thin / R1_thick / R2_puppet / heatmap),
+`hands` / `face` / `feet` toggles, `redetect`, `resolution`, hidden `pose_json`.
+**Outputs:** `skeleton` (IMAGE) · `keypoints` (POSE_KEYPOINT).
+
+### Use
+1. Connect a reference photo to `image`, **Queue** once (`redetect` on) — detects the pose, fills the canvas,
+   paints the styled skeleton on the node.
+2. **Mute** the node's re-detect (toggle `redetect` off — auto-set when you drag) and drag body joints in the
+   mini-canvas.
+3. **Queue** — only this node + downstream re-run (detection is skipped). Output feeds `ImageScale → VAEEncode`.
+
+### Styles & toggles
+`R0_thin` = DWPose-style thin skeleton (matches typical pose-LoRA training); `R1_thick` / `R2_puppet` =
+thicker limbs; `heatmap` = gaussian joint blobs. `hands`/`face`/`feet` add/remove those parts (R1/R2/heatmap
+are body-only; `feet` adds foot joints). No color picker — the palette is the fixed DWPose palette.
+
+Pure-Python renderer (`pose_render.py`) is the single source of truth and is unit-tested headless.

comfyui/ComfyUI-anima-pose-control/__init__.py

@@ -0,0 +1,6 @@
+from .node import AnimaPoseControl
+
+NODE_CLASS_MAPPINGS = {"AnimaPoseControl": AnimaPoseControl}
+NODE_DISPLAY_NAME_MAPPINGS = {"AnimaPoseControl": "Anima Pose Control (detect+edit+style)"}
+WEB_DIRECTORY = "./js"
+__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS", "WEB_DIRECTORY"]

comfyui/ComfyUI-anima-pose-control/detect.py

@@ -0,0 +1,127 @@
+"""DWPose detection wrapper. Uses rtmlib Wholebody — the same model + 133-keypoint convention
+v2 trained on (render_skeletons.py / eval used the identical call). Injectable for tests.
+
+Wholebody-133: body 0-16, feet 17-22, face 23-90, L-hand 91-111, R-hand 112-132.
+
+On first use rtmlib's Wholebody (mode="balanced") needs two ONNX files. rtmlib would fetch them
+from download.openmmlab.com, which is slow or blocked on many networks ("getaddrinfo failed").
+We mirror them on the model's own HF repo and pre-seed rtmlib's cache from there, so the node
+works for anyone who can reach huggingface.co. openmmlab stays a silent fallback."""
+import os
+import numpy as np
+
+
+class NoPersonError(RuntimeError):
+    pass
+
+
+_POSE = None
+
+# rtmlib Wholebody(mode="balanced") checkpoints, mirrored on the model repo to dodge the
+# flaky download.openmmlab.com host. Filenames MUST match the names rtmlib looks for in cache.
+_HF_REPO = "Claquasse/Anima-Control-Pose"
+_MODELS = (
+    "yolox_m_8xb8-300e_humanart-c2c7a14a.onnx",
+    "rtmw-dw-x-l_simcc-cocktail14_270e-256x192_20231122.onnx",
+)
+
+
+def _rtmlib_ckpt_dir():
+    try:
+        from rtmlib.tools.file import _get_rtmhub_dir
+        return os.path.join(_get_rtmhub_dir(), "checkpoints")
+    except Exception:
+        base = os.getenv("XDG_CACHE_HOME") or os.path.expanduser("~/.cache")
+        return os.path.join(os.path.expanduser(base), "rtmlib", "hub", "checkpoints")
+
+
+# huggingface.co is often blocked/throttled in mainland China; hf-mirror.com mirrors it and is
+# reachable there. We try the Hub first (honoring HF_ENDPOINT if the user set one), then the mirror.
+_MIRROR = "https://hf-mirror.com"
+
+
+def _looks_real(path):
+    # a real .onnx is tens-to-hundreds of MB; an HTML error page or partial is tiny.
+    return os.path.exists(path) and os.path.getsize(path) > 1_000_000
+
+
+def _fetch(name, dst):
+    """Download detector/<name> to dst, trying the HF Hub then the China-friendly mirror.
+    Returns True only on a plausibly-complete file (guards against caching an error page)."""
+    import shutil
+    try:
+        from huggingface_hub import hf_hub_download
+        src = hf_hub_download(repo_id=_HF_REPO, filename="detector/" + name)
+        shutil.copyfile(src, dst)
+        if _looks_real(dst):
+            return True
+    except Exception:
+        pass
+    try:
+        import urllib.request
+        url = "%s/%s/resolve/main/detector/%s" % (_MIRROR, _HF_REPO, name)
+        with urllib.request.urlopen(url, timeout=120) as r, open(dst, "wb") as f:
+            shutil.copyfileobj(r, f)
+        if _looks_real(dst):
+            return True
+    except Exception:
+        pass
+    if os.path.exists(dst) and not _looks_real(dst):
+        try:
+            os.remove(dst)            # never leave a partial/HTML stub behind
+        except OSError:
+            pass
+    return False
+
+
+def _ensure_models():
+    """Pre-seed rtmlib's checkpoint cache from the HF model repo (or its mirror) so it never hits
+    download.openmmlab.com. Best-effort: on failure, rtmlib falls back to its own download."""
+    ckpt_dir = _rtmlib_ckpt_dir()
+    # A leftover <base>.zip means an earlier download failed mid-way (openmmlab returned a
+    # partial file or an HTML error page). rtmlib would then try to extract it and crash with
+    # `BadZipFile: File is not a zip file` on every run, even when the .onnx is present, because
+    # its skip-download path only triggers when no .zip exists. Remove ours so the seeded .onnx
+    # is used instead.
+    for name in _MODELS:
+        zp = os.path.join(ckpt_dir, os.path.splitext(name)[0] + ".zip")
+        if os.path.exists(zp):
+            try:
+                os.remove(zp)
+            except OSError:
+                pass
+    missing = [m for m in _MODELS if not os.path.exists(os.path.join(ckpt_dir, m))]
+    if not missing:
+        return
+    os.makedirs(ckpt_dir, exist_ok=True)
+    for name in missing:
+        if not _fetch(name, os.path.join(ckpt_dir, name)):
+            print("[AnimaPoseControl] could not fetch %s from Hugging Face or its mirror. "
+                  "rtmlib will try download.openmmlab.com instead; if that fails, see the model "
+                  "card for the offline install." % name)
+
+
+def _default_estimator(resolution):
+    """Lazily build the rtmlib Wholebody estimator (ComfyUI venv). Returns a callable
+    img_rgb -> (kp[N,133,2], sc[N,133]) in pixel coords on a `resolution`-square canvas."""
+    global _POSE
+    if _POSE is None:
+        _ensure_models()
+        from rtmlib import Wholebody
+        _POSE = Wholebody(to_openpose=False, mode="balanced", backend="onnxruntime", device="cpu")
+    pose = _POSE
+    res = int(resolution)
+
+    def run(img):
+        import cv2
+        bgr = cv2.cvtColor(cv2.resize(img, (res, res)), cv2.COLOR_RGB2BGR)
+        return pose(bgr)            # (N,133,2), (N,133)
+    return run
+
+
+def detect(image_rgb, resolution, estimator=None):
+    est = _default_estimator(resolution) if estimator is None else estimator
+    kp_all, sc_all = est(image_rgb)
+    if kp_all is None or len(kp_all) == 0:
+        raise NoPersonError("no person detected")
+    return np.asarray(kp_all[0], float), np.asarray(sc_all[0], float)

comfyui/ComfyUI-anima-pose-control/js/anima_pose_control.js

@@ -0,0 +1,16 @@
+// js/anima_pose_control.js — Anima Pose Control: hide the machine-managed pose_json text box.
+// (Joint-drag editing removed for now; the node shows the styled skeleton via ComfyUI's image preview.)
+import { app } from "../../scripts/app.js";
+
+app.registerExtension({
+  name: "anima.pose.control",
+  async beforeRegisterNodeDef(nodeType, nodeData) {
+    if (nodeData.name !== "AnimaPoseControl") return;
+    const onNodeCreated = nodeType.prototype.onNodeCreated;
+    nodeType.prototype.onNodeCreated = function () {
+      onNodeCreated?.apply(this, arguments);
+      const pj = this.widgets?.find((w) => w.name === "pose_json");
+      if (pj) { pj.hidden = true; pj.computeSize = () => [0, -4]; pj.type = "hidden_pose_json"; }
+    };
+  },
+});

comfyui/ComfyUI-anima-pose-control/keypoints.py

@@ -0,0 +1,21 @@
+"""Codec between the JS canvas pose JSON and numpy keypoint arrays, plus body-17 edit merge."""
+import json
+import numpy as np
+
+
+def to_json(kp, sc, canvas):
+    kp = np.asarray(kp, float); sc = np.asarray(sc, float)
+    pts = [[float(kp[i][0]), float(kp[i][1]), float(sc[i])] for i in range(133)]
+    return json.dumps({"canvas": int(canvas), "points": pts})
+
+
+def from_json(s):
+    d = json.loads(s)
+    pts = np.asarray(d["points"], float)            # (133,3)
+    return pts[:, :2].copy(), pts[:, 2].copy(), int(d["canvas"])
+
+
+def merge_body(full_kp, full_sc, body_kp):
+    out = np.asarray(full_kp, float).copy()
+    out[:17] = np.asarray(body_kp, float)
+    return out

comfyui/ComfyUI-anima-pose-control/node.py

@@ -0,0 +1,70 @@
+"""AnimaPoseControl ComfyUI node: detect -> edit (JS) -> styled skeleton -> control IMAGE."""
+import numpy as np
+
+try:                                   # keep the helper importable without the package context (tests)
+    from .pose_render import render, STYLES
+    from .keypoints import from_json, to_json
+    from .detect import detect
+except ImportError:                    # flat import when running tests inside the package dir
+    from pose_render import render, STYLES
+    from keypoints import from_json, to_json
+    from detect import detect
+
+
+def compose(image, pose_json, style, hands, face, feet, redetect, resolution, estimator=None):
+    """Pure branch logic. Returns (skeleton uint8 [H,W,3] RGB, pose_json_out str)."""
+    if redetect and image is not None:
+        kp, sc = detect(image, int(resolution), estimator=estimator)
+    elif pose_json and pose_json.strip():
+        kp, sc, _ = from_json(pose_json)
+    elif image is not None:
+        kp, sc = detect(image, int(resolution), estimator=estimator)   # no pose yet -> detect
+    else:
+        raise ValueError("AnimaPoseControl: no pose to render — connect an image, or Queue once with redetect on first")
+    img = render(kp, sc, style, int(resolution), hands=hands, face=face, feet=feet)
+    return img, to_json(kp, sc, int(resolution))
+
+
+class AnimaPoseControl:
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {
+            "required": {
+                "style": (list(STYLES),),
+                "hands": ("BOOLEAN", {"default": True}),
+                "face": ("BOOLEAN", {"default": True}),
+                "feet": ("BOOLEAN", {"default": True}),
+                "redetect": ("BOOLEAN", {"default": True}),
+                "resolution": ("INT", {"default": 768, "min": 256, "max": 2048, "step": 64}),
+                "pose_json": ("STRING", {"multiline": True, "default": ""}),
+            },
+            "optional": {"image": ("IMAGE",)},
+        }
+
+    RETURN_TYPES = ("IMAGE", "POSE_KEYPOINT")
+    RETURN_NAMES = ("skeleton", "keypoints")
+    FUNCTION = "run"
+    OUTPUT_NODE = True
+    CATEGORY = "Anima/pose"
+
+    def run(self, style, hands, face, feet, redetect, resolution, pose_json, image=None):
+        import torch
+        img_in = None
+        if image is not None:
+            img_in = (image[0].cpu().numpy() * 255).astype(np.uint8)   # ComfyUI IMAGE -> HWC uint8
+        skel, pj = compose(img_in, pose_json, style, hands, face, feet, redetect, resolution)
+        t = torch.from_numpy(skel.astype(np.float32) / 255.0)[None, ...]
+        import json
+        kpts = json.loads(pj)
+        # Show the styled render inline via ComfyUI's standard image preview.
+        return {"ui": {"images": _save_preview(skel), "pose_json": [pj]}, "result": (t, kpts)}
+
+
+def _save_preview(skel_uint8):
+    """Write the styled skeleton to ComfyUI's temp dir; return a /view descriptor for the JS to load."""
+    import folder_paths, os, uuid
+    from PIL import Image
+    d = folder_paths.get_temp_directory(); os.makedirs(d, exist_ok=True)
+    fn = "anima_pose_%s.png" % uuid.uuid4().hex[:8]
+    Image.fromarray(skel_uint8).save(os.path.join(d, fn))
+    return [{"filename": fn, "subfolder": "", "type": "temp"}]

comfyui/ComfyUI-anima-pose-control/pose_render.py

@@ -0,0 +1,62 @@
+"""Styled pose-skeleton renderer for AnimaPoseControl. Pure numpy + cv2 (no torch).
+Single source of truth for the predefined skeleton looks.
+Wholebody-133: body 0-16, feet 17-22, face 23-90, L-hand 91-111, R-hand 112-132."""
+import numpy as np
+import cv2
+
+BODY = list(range(17)); FEET = list(range(17, 23)); FACE = list(range(23, 91)); HAND = list(range(91, 133))
+COCO_LIMBS = [(5, 7), (7, 9), (6, 8), (8, 10), (11, 13), (13, 15), (12, 14), (14, 16),
+              (5, 6), (11, 12), (5, 11), (6, 12), (0, 5), (0, 6)]
+_PAL = [(255, 0, 0), (255, 128, 0), (255, 255, 0), (128, 255, 0), (0, 255, 0), (0, 255, 128), (0, 255, 255),
+        (0, 128, 255), (0, 0, 255), (128, 0, 255), (255, 0, 255), (255, 0, 128), (180, 180, 180), (220, 220, 220)]
+LIMB_COLOR = {l: _PAL[i % len(_PAL)] for i, l in enumerate(COCO_LIMBS)}
+JOINT_COLOR = [_PAL[i % len(_PAL)] for i in range(17)]
+STYLES = ("R0_thin", "R1_thick", "R2_puppet", "heatmap")
+_W = {"R0_thin": {"stick": 2, "joint": 3}, "R1_thick": {"stick": 8, "joint": 6}, "R2_puppet": {"stick": 22, "joint": 11}, "heatmap": {"stick": 2, "joint": 3}}
+
+
+def _toggle(sc, hands, face, feet):
+    sc = np.asarray(sc, float).copy()
+    if not hands: sc[HAND] = 0.0
+    if not face: sc[FACE] = 0.0
+    if not feet: sc[FEET] = 0.0
+    return sc
+
+
+def render(kp, sc, style, resolution, hands=True, face=True, feet=True, kpt_thr=0.3):
+    if style not in STYLES:
+        raise ValueError(f"unknown style {style!r}")
+    H = W = int(resolution)
+    kp = np.asarray(kp, float); sc = _toggle(sc, hands, face, feet)
+    if style == "heatmap":
+        return _heatmap(kp, sc, H, W, feet, kpt_thr)
+    cfg = _W[style]; cv = np.zeros((H, W, 3), np.uint8)
+    for a, b in COCO_LIMBS:
+        if sc[a] >= kpt_thr and sc[b] >= kpt_thr:
+            cv2.line(cv, (int(kp[a][0]), int(kp[a][1])), (int(kp[b][0]), int(kp[b][1])), LIMB_COLOR[(a, b)], cfg["stick"])
+    for i in BODY:
+        if sc[i] >= kpt_thr:
+            cv2.circle(cv, (int(kp[i][0]), int(kp[i][1])), cfg["joint"], (255, 255, 255), -1)
+    for i in FEET:
+        if sc[i] >= kpt_thr:
+            cv2.circle(cv, (int(kp[i][0]), int(kp[i][1])), max(1, cfg["joint"] // 2), (255, 255, 255), -1)
+    if style == "R0_thin":              # only R0 draws hands + face (DWPose-style small visible dots)
+        for i in HAND:
+            if sc[i] >= kpt_thr:
+                cv2.circle(cv, (int(kp[i][0]), int(kp[i][1])), 2, (0, 255, 255), -1)
+        for i in FACE:
+            if sc[i] >= kpt_thr:
+                cv2.circle(cv, (int(kp[i][0]), int(kp[i][1])), 2, (255, 255, 255), -1)
+    return cv
+
+
+def _heatmap(kp, sc, H, W, feet, thr, sig=11):
+    acc = np.zeros((H, W, 3), np.float32)
+    for i in BODY + (FEET if feet else []):
+        if sc[i] >= thr:
+            x, y = int(round(kp[i][0])), int(round(kp[i][1]))
+            if 0 <= x < W and 0 <= y < H:
+                b = np.zeros((H, W), np.float32); b[y, x] = 1.0
+                b = cv2.GaussianBlur(b, (0, 0), sig); b /= (b.max() + 1e-9)
+                acc += b[..., None] * np.array(JOINT_COLOR[i % 17], np.float32)
+    return np.clip(acc, 0, 255).astype(np.uint8)

comfyui/ComfyUI-anima-pose-control/requirements.txt

@@ -0,0 +1,8 @@
+# ComfyUI-anima-pose-control runtime dependencies.
+# ComfyUI already provides torch, Pillow and numpy; these are the extras the node needs.
+# ComfyUI-Manager installs this automatically on first load.
+rtmlib
+opencv-python
+onnxruntime
+numpy
+huggingface_hub

comfyui/anima_control_lora/__init__.py

@@ -0,0 +1,73 @@
+"""ComfyUI node: fuse a trained control-LoRA's ControlEmbedder into the Cosmos DiT forward.
+Wraps the DIFFUSION_MODEL forward (comfy/model_patcher.py add_wrapper_with_key) to add the
+control contribution at the patch-embed output; LoRA deltas on existing weights are applied
+separately via LoraLoaderModelOnly upstream of this node."""
+
+
+class AnimaControlApply:
+    @classmethod
+    def INPUT_TYPES(cls):
+        return {"required": {
+            "model": ("MODEL",),
+            "control_latent": ("LATENT",),            # VAE-encoded skeleton (use a VAEEncode node)
+            "control_embedder_path": ("STRING", {"default": "adapter_model.safetensors"}),
+            "strength": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 2.0, "step": 0.05}),
+        }}
+
+    RETURN_TYPES = ("MODEL",)
+    FUNCTION = "apply"
+    CATEGORY = "AnimaControl"
+
+    def apply(self, model, control_latent, control_embedder_path, strength):
+        import safetensors.torch as st
+        from .control_embedder import ControlEmbedder
+        dit = model.model.diffusion_model
+        p = next(dit.parameters())
+        embedder = ControlEmbedder(in_channels=16, model_channels=getattr(dit, "model_channels", 2048))
+        # Training saves the embedder INSIDE the adapter file as `diffusion_model.control_embedder.proj.*`
+        # (saver keeps `control_embedder.proj.*`, base save_adapter prepends `diffusion_model.`). Strip up
+        # to and including `control_embedder.`, and fail loudly if nothing matches — otherwise the embedder
+        # would stay zero-init and control would be a silent no-op.
+        # Resolve a bare filename against ComfyUI's loras/ dir so the same downloaded
+        # adapter_model.safetensors feeds both this node and LoraLoaderModelOnly; absolute
+        # paths are used as-is (the training/eval scripts pass those).
+        import os
+        cep = control_embedder_path
+        if not os.path.isabs(cep):
+            try:
+                import folder_paths
+                cep = folder_paths.get_full_path("loras", cep) or cep
+            except Exception:
+                pass
+        sd = {k.split("control_embedder.", 1)[1]: v
+              for k, v in st.load_file(cep).items() if "control_embedder." in k}
+        assert sd, f"no control_embedder.* keys in {cep} (expected diffusion_model.control_embedder.proj.*)"
+        embedder.load_state_dict(sd, strict=False)
+        embedder.eval().to(p.device, p.dtype)
+        ctrl = control_latent["samples"]
+
+        def wrapper(executor, *args, **kwargs):
+            # Reproduce training-time fusion (ControlInitialLayer._fuse_control): add the
+            # ControlEmbedder output to the x_embedder output via a temporary forward hook,
+            # removed after the call so it never leaks onto the shared DiT. strength=0 zeroes
+            # control_tokens -> output unchanged (the base regression).
+            control_tokens = strength * embedder(ctrl.to(p.device, p.dtype))
+
+            def add_control(_module, _inputs, output):
+                # match output's device AND dtype — at apply() the DiT may still be on CPU (ComfyUI
+                # lazy-loads), so control_tokens can be built on CPU while output is on cuda at sampling.
+                return output + control_tokens.to(output.device, output.dtype)
+
+            handle = dit.x_embedder.register_forward_hook(add_control)
+            try:
+                return executor(*args, **kwargs)
+            finally:
+                handle.remove()
+
+        m = model.clone()
+        m.add_wrapper_with_key("diffusion_model", "anima_control", wrapper)
+        return (m,)
+
+
+NODE_CLASS_MAPPINGS = {"AnimaControlApply": AnimaControlApply}
+NODE_DISPLAY_NAME_MAPPINGS = {"AnimaControlApply": "Anima Control Apply"}

comfyui/anima_control_lora/control_embedder.py

@@ -0,0 +1,29 @@
+import torch.nn as nn
+from einops import rearrange
+
+
+class ControlEmbedder(nn.Module):
+    """Zero-init patch-embed for the control (pose-skeleton) latent.
+
+    Mirrors the base PatchEmbed (spatial_patch_size=2, temporal_patch_size=1) but on the
+    raw 16-channel control latent (no padding-mask channel). Zero-initialized so it adds
+    nothing until trained (ControlNet zero-init). Output matches prepare_embedded_sequence's
+    x_B_T_H_W_D: (B, T, H/2, W/2, model_channels).
+    """
+
+    def __init__(self, in_channels=16, spatial_patch_size=2, temporal_patch_size=1, model_channels=2048):
+        super().__init__()
+        self.spatial_patch_size = spatial_patch_size
+        self.temporal_patch_size = temporal_patch_size
+        in_features = in_channels * spatial_patch_size * spatial_patch_size * temporal_patch_size
+        self.proj = nn.Linear(in_features, model_channels)
+        nn.init.zeros_(self.proj.weight)
+        nn.init.zeros_(self.proj.bias)
+
+    def forward(self, control_B_C_T_H_W):
+        x = rearrange(
+            control_B_C_T_H_W,
+            "b c (t r) (h m) (w n) -> b t h w (c r m n)",
+            r=self.temporal_patch_size, m=self.spatial_patch_size, n=self.spatial_patch_size,
+        )
+        return self.proj(x)

detector/rtmw-dw-x-l_simcc-cocktail14_270e-256x192_20231122.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a9bc17c13ff0a1e37507f45a037a21bc410cd08374b0e664c57d0080843aa58
+size 228708578

detector/yolox_m_8xb8-300e_humanart-c2c7a14a.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3dea6513388889f0fff4b77bf7a26013600321b9eb9ceb0e9a400a82572f5f23
+size 101400344

pose_control.json

@@ -0,0 +1,1084 @@
+{
+  "id": "release",
+  "revision": 0,
+  "last_node_id": 16,
+  "last_link_id": 17,
+  "nodes": [
+    {
+      "id": 1,
+      "type": "UNETLoader",
+      "pos": [
+        40,
+        40
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 1,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "unet_name",
+          "name": "unet_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "unet_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "weight_dtype",
+          "name": "weight_dtype",
+          "type": "COMBO",
+          "widget": {
+            "name": "weight_dtype"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            1
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "UNETLoader"
+      },
+      "widgets_values": [
+        "anima-base-v1.0.safetensors",
+        "default"
+      ]
+    },
+    {
+      "id": 2,
+      "type": "LoraLoaderModelOnly",
+      "pos": [
+        40,
+        180
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 2,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 1
+        },
+        {
+          "localized_name": "lora_name",
+          "name": "lora_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "lora_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "strength_model",
+          "name": "strength_model",
+          "type": "FLOAT",
+          "widget": {
+            "name": "strength_model"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            8
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "LoraLoaderModelOnly"
+      },
+      "widgets_values": [
+        "anima_pose_preview2.safetensors",
+        1
+      ]
+    },
+    {
+      "id": 3,
+      "type": "CLIPLoader",
+      "pos": [
+        40,
+        320
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 3,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip_name",
+          "name": "clip_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "clip_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "type",
+          "name": "type",
+          "type": "COMBO",
+          "widget": {
+            "name": "type"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "device",
+          "name": "device",
+          "shape": 7,
+          "type": "COMBO",
+          "widget": {
+            "name": "device"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CLIP",
+          "name": "CLIP",
+          "type": "CLIP",
+          "slot_index": 0,
+          "links": [
+            2,
+            3
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPLoader"
+      },
+      "widgets_values": [
+        "qwen_3_06b_base.safetensors",
+        "stable_diffusion",
+        "default"
+      ]
+    },
+    {
+      "id": 4,
+      "type": "VAELoader",
+      "pos": [
+        40,
+        450
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 4,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "vae_name",
+          "name": "vae_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "vae_name"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "VAE",
+          "name": "VAE",
+          "type": "VAE",
+          "slot_index": 0,
+          "links": [
+            7,
+            16
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAELoader"
+      },
+      "widgets_values": [
+        "qwen_image_vae.safetensors"
+      ]
+    },
+    {
+      "id": 5,
+      "type": "CLIPTextEncode",
+      "pos": [
+        40,
+        560
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 5,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip",
+          "name": "clip",
+          "type": "CLIP",
+          "link": 2
+        },
+        {
+          "localized_name": "text",
+          "name": "text",
+          "type": "STRING",
+          "widget": {
+            "name": "text"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CONDITIONING",
+          "name": "CONDITIONING",
+          "type": "CONDITIONING",
+          "slot_index": 0,
+          "links": [
+            12
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPTextEncode"
+      },
+      "widgets_values": [
+        "newest, masterpiece, best quality, score_9, safe\n\n1girl, solo, simple background"
+      ],
+      "title": "prompt"
+    },
+    {
+      "id": 6,
+      "type": "CLIPTextEncode",
+      "pos": [
+        40,
+        760
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 6,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip",
+          "name": "clip",
+          "type": "CLIP",
+          "link": 3
+        },
+        {
+          "localized_name": "text",
+          "name": "text",
+          "type": "STRING",
+          "widget": {
+            "name": "text"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CONDITIONING",
+          "name": "CONDITIONING",
+          "type": "CONDITIONING",
+          "slot_index": 0,
+          "links": [
+            13
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPTextEncode"
+      },
+      "widgets_values": [
+        "worst quality, low quality, score_1, score_2, blurry, jpeg artifacts"
+      ],
+      "title": "negative"
+    },
+    {
+      "id": 7,
+      "type": "EmptyLatentImage",
+      "pos": [
+        40,
+        960
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 7,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "width",
+          "name": "width",
+          "type": "INT",
+          "widget": {
+            "name": "width"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "height",
+          "name": "height",
+          "type": "INT",
+          "widget": {
+            "name": "height"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "batch_size",
+          "name": "batch_size",
+          "type": "INT",
+          "widget": {
+            "name": "batch_size"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            14
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "EmptyLatentImage"
+      },
+      "widgets_values": [
+        768,
+        768,
+        1
+      ]
+    },
+    {
+      "id": 8,
+      "type": "LoadImage",
+      "pos": [
+        380,
+        120
+      ],
+      "size": [
+        300,
+        314
+      ],
+      "flags": {},
+      "order": 8,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "image",
+          "name": "image",
+          "type": "COMBO",
+          "widget": {
+            "name": "image"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "choose file to upload",
+          "name": "upload",
+          "type": "IMAGEUPLOAD",
+          "widget": {
+            "name": "upload"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            4
+          ]
+        },
+        {
+          "localized_name": "MASK",
+          "name": "MASK",
+          "type": "MASK",
+          "slot_index": 1,
+          "links": []
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "LoadImage"
+      },
+      "widgets_values": [
+        "cmp_source.png",
+        "image"
+      ],
+      "title": "reference PHOTO"
+    },
+    {
+      "id": 9,
+      "type": "AnimaPoseControl",
+      "pos": [
+        680,
+        120
+      ],
+      "size": [
+        260,
+        200
+      ],
+      "flags": {},
+      "order": 9,
+      "mode": 0,
+      "properties": {
+        "Node name for S&R": "AnimaPoseControl"
+      },
+      "widgets_values": [
+        "R0_thin",
+        true,
+        true,
+        true,
+        true,
+        768,
+        ""
+      ],
+      "inputs": [
+        {
+          "name": "image",
+          "type": "IMAGE",
+          "link": 4
+        }
+      ],
+      "outputs": [
+        {
+          "name": "skeleton",
+          "type": "IMAGE",
+          "links": [
+            5
+          ],
+          "slot_index": 0
+        },
+        {
+          "name": "keypoints",
+          "type": "POSE_KEYPOINT",
+          "links": [],
+          "slot_index": 1
+        }
+      ],
+      "title": "POSE: detect + style",
+      "color": "#533",
+      "bgcolor": "#223"
+    },
+    {
+      "id": 10,
+      "type": "ImageScale",
+      "pos": [
+        980,
+        120
+      ],
+      "size": [
+        300,
+        130
+      ],
+      "flags": {},
+      "order": 10,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "image",
+          "name": "image",
+          "type": "IMAGE",
+          "link": 5
+        },
+        {
+          "localized_name": "upscale_method",
+          "name": "upscale_method",
+          "type": "COMBO",
+          "widget": {
+            "name": "upscale_method"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "width",
+          "name": "width",
+          "type": "INT",
+          "widget": {
+            "name": "width"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "height",
+          "name": "height",
+          "type": "INT",
+          "widget": {
+            "name": "height"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "crop",
+          "name": "crop",
+          "type": "COMBO",
+          "widget": {
+            "name": "crop"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            6
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "ImageScale"
+      },
+      "widgets_values": [
+        "nearest-exact",
+        768,
+        768,
+        "disabled"
+      ]
+    },
+    {
+      "id": 11,
+      "type": "VAEEncode",
+      "pos": [
+        1280,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 11,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "pixels",
+          "name": "pixels",
+          "type": "IMAGE",
+          "link": 6
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 7
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            9
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEEncode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 12,
+      "type": "AnimaControlApply",
+      "pos": [
+        1580,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 12,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 8
+        },
+        {
+          "localized_name": "control_latent",
+          "name": "control_latent",
+          "type": "LATENT",
+          "link": 9
+        },
+        {
+          "localized_name": "control_embedder_path",
+          "name": "control_embedder_path",
+          "type": "STRING",
+          "widget": {
+            "name": "control_embedder_path"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "strength",
+          "name": "strength",
+          "type": "FLOAT",
+          "widget": {
+            "name": "strength"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            10
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "AnimaControlApply"
+      },
+      "widgets_values": [
+        "anima_pose_preview2.safetensors",
+        1
+      ]
+    },
+    {
+      "id": 13,
+      "type": "CFGZeroStar",
+      "pos": [
+        1880,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 13,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 10
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "patched_model",
+          "name": "patched_model",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            11
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CFGZeroStar"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 14,
+      "type": "KSampler",
+      "pos": [
+        2180,
+        120
+      ],
+      "size": [
+        300,
+        262
+      ],
+      "flags": {},
+      "order": 14,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 11
+        },
+        {
+          "localized_name": "positive",
+          "name": "positive",
+          "type": "CONDITIONING",
+          "link": 12
+        },
+        {
+          "localized_name": "negative",
+          "name": "negative",
+          "type": "CONDITIONING",
+          "link": 13
+        },
+        {
+          "localized_name": "latent_image",
+          "name": "latent_image",
+          "type": "LATENT",
+          "link": 14
+        },
+        {
+          "localized_name": "seed",
+          "name": "seed",
+          "type": "INT",
+          "widget": {
+            "name": "seed"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "steps",
+          "name": "steps",
+          "type": "INT",
+          "widget": {
+            "name": "steps"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "cfg",
+          "name": "cfg",
+          "type": "FLOAT",
+          "widget": {
+            "name": "cfg"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "sampler_name",
+          "name": "sampler_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "sampler_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "scheduler",
+          "name": "scheduler",
+          "type": "COMBO",
+          "widget": {
+            "name": "scheduler"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "denoise",
+          "name": "denoise",
+          "type": "FLOAT",
+          "widget": {
+            "name": "denoise"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            15
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "KSampler"
+      },
+      "widgets_values": [
+        2024,
+        "fixed",
+        30,
+        4,
+        "er_sde",
+        "simple",
+        1
+      ]
+    },
+    {
+      "id": 15,
+      "type": "VAEDecode",
+      "pos": [
+        2480,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 15,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "samples",
+          "name": "samples",
+          "type": "LATENT",
+          "link": 15
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 16
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            17
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEDecode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 16,
+      "type": "SaveImage",
+      "pos": [
+        2780,
+        120
+      ],
+      "size": [
+        300,
+        270
+      ],
+      "flags": {},
+      "order": 16,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "images",
+          "name": "images",
+          "type": "IMAGE",
+          "link": 17
+        },
+        {
+          "localized_name": "filename_prefix",
+          "name": "filename_prefix",
+          "type": "STRING",
+          "widget": {
+            "name": "filename_prefix"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [],
+      "properties": {
+        "Node name for S&R": "SaveImage"
+      },
+      "widgets_values": [
+        "anima_pose_control"
+      ],
+      "title": "GENERATION"
+    }
+  ],
+  "links": [
+    [
+      1,
+      1,
+      0,
+      2,
+      0,
+      "MODEL"
+    ],
+    [
+      2,
+      3,
+      0,
+      5,
+      0,
+      "CLIP"
+    ],
+    [
+      3,
+      3,
+      0,
+      6,
+      0,
+      "CLIP"
+    ],
+    [
+      4,
+      8,
+      0,
+      9,
+      0,
+      "IMAGE"
+    ],
+    [
+      5,
+      9,
+      0,
+      10,
+      0,
+      "IMAGE"
+    ],
+    [
+      6,
+      10,
+      0,
+      11,
+      0,
+      "IMAGE"
+    ],
+    [
+      7,
+      4,
+      0,
+      11,
+      1,
+      "VAE"
+    ],
+    [
+      8,
+      2,
+      0,
+      12,
+      0,
+      "MODEL"
+    ],
+    [
+      9,
+      11,
+      0,
+      12,
+      1,
+      "LATENT"
+    ],
+    [
+      10,
+      12,
+      0,
+      13,
+      0,
+      "MODEL"
+    ],
+    [
+      11,
+      13,
+      0,
+      14,
+      0,
+      "MODEL"
+    ],
+    [
+      12,
+      5,
+      0,
+      14,
+      1,
+      "CONDITIONING"
+    ],
+    [
+      13,
+      6,
+      0,
+      14,
+      2,
+      "CONDITIONING"
+    ],
+    [
+      14,
+      7,
+      0,
+      14,
+      3,
+      "LATENT"
+    ],
+    [
+      15,
+      14,
+      0,
+      15,
+      0,
+      "LATENT"
+    ],
+    [
+      16,
+      4,
+      0,
+      15,
+      1,
+      "VAE"
+    ],
+    [
+      17,
+      15,
+      0,
+      16,
+      0,
+      "IMAGE"
+    ]
+  ],
+  "groups": [],
+  "config": {},
+  "extra": {
+    "ds": {
+      "scale": 1.2284597357367357,
+      "offset": [
+        -896.6673282988922,
+        -2098.6519356376084
+      ]
+    }
+  },
+  "version": 0.4
+}

pose_control_demo.json

@@ -0,0 +1,1469 @@
+{
+  "id": "release",
+  "revision": 0,
+  "last_node_id": 21,
+  "last_link_id": 26,
+  "nodes": [
+    {
+      "id": 1,
+      "type": "UNETLoader",
+      "pos": [
+        40,
+        40
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 1,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "unet_name",
+          "name": "unet_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "unet_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "weight_dtype",
+          "name": "weight_dtype",
+          "type": "COMBO",
+          "widget": {
+            "name": "weight_dtype"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            1,
+            18
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "UNETLoader"
+      },
+      "widgets_values": [
+        "anima-base-v1.0.safetensors",
+        "default"
+      ]
+    },
+    {
+      "id": 2,
+      "type": "LoraLoaderModelOnly",
+      "pos": [
+        40,
+        180
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 2,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 1
+        },
+        {
+          "localized_name": "lora_name",
+          "name": "lora_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "lora_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "strength_model",
+          "name": "strength_model",
+          "type": "FLOAT",
+          "widget": {
+            "name": "strength_model"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            8
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "LoraLoaderModelOnly"
+      },
+      "widgets_values": [
+        "anima_pose_preview2.safetensors",
+        1
+      ]
+    },
+    {
+      "id": 3,
+      "type": "CLIPLoader",
+      "pos": [
+        40,
+        320
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 3,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip_name",
+          "name": "clip_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "clip_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "type",
+          "name": "type",
+          "type": "COMBO",
+          "widget": {
+            "name": "type"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "device",
+          "name": "device",
+          "shape": 7,
+          "type": "COMBO",
+          "widget": {
+            "name": "device"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CLIP",
+          "name": "CLIP",
+          "type": "CLIP",
+          "slot_index": 0,
+          "links": [
+            2,
+            3
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPLoader"
+      },
+      "widgets_values": [
+        "qwen_3_06b_base.safetensors",
+        "stable_diffusion",
+        "default"
+      ]
+    },
+    {
+      "id": 4,
+      "type": "VAELoader",
+      "pos": [
+        40,
+        450
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 4,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "vae_name",
+          "name": "vae_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "vae_name"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "VAE",
+          "name": "VAE",
+          "type": "VAE",
+          "slot_index": 0,
+          "links": [
+            7,
+            16,
+            24
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAELoader"
+      },
+      "widgets_values": [
+        "qwen_image_vae.safetensors"
+      ]
+    },
+    {
+      "id": 5,
+      "type": "CLIPTextEncode",
+      "pos": [
+        40,
+        560
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 5,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip",
+          "name": "clip",
+          "type": "CLIP",
+          "link": 2
+        },
+        {
+          "localized_name": "text",
+          "name": "text",
+          "type": "STRING",
+          "widget": {
+            "name": "text"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CONDITIONING",
+          "name": "CONDITIONING",
+          "type": "CONDITIONING",
+          "slot_index": 0,
+          "links": [
+            12,
+            20
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPTextEncode"
+      },
+      "widgets_values": [
+        "newest, masterpiece, best quality, score_9, safe\n\n1girl, solo, simple background"
+      ],
+      "title": "prompt"
+    },
+    {
+      "id": 6,
+      "type": "CLIPTextEncode",
+      "pos": [
+        40,
+        760
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 6,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip",
+          "name": "clip",
+          "type": "CLIP",
+          "link": 3
+        },
+        {
+          "localized_name": "text",
+          "name": "text",
+          "type": "STRING",
+          "widget": {
+            "name": "text"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CONDITIONING",
+          "name": "CONDITIONING",
+          "type": "CONDITIONING",
+          "slot_index": 0,
+          "links": [
+            13,
+            21
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPTextEncode"
+      },
+      "widgets_values": [
+        "worst quality, low quality, score_1, score_2, blurry, jpeg artifacts"
+      ],
+      "title": "negative"
+    },
+    {
+      "id": 7,
+      "type": "EmptyLatentImage",
+      "pos": [
+        40,
+        960
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 7,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "width",
+          "name": "width",
+          "type": "INT",
+          "widget": {
+            "name": "width"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "height",
+          "name": "height",
+          "type": "INT",
+          "widget": {
+            "name": "height"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "batch_size",
+          "name": "batch_size",
+          "type": "INT",
+          "widget": {
+            "name": "batch_size"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            14,
+            22
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "EmptyLatentImage"
+      },
+      "widgets_values": [
+        768,
+        768,
+        1
+      ]
+    },
+    {
+      "id": 8,
+      "type": "LoadImage",
+      "pos": [
+        380,
+        120
+      ],
+      "size": [
+        300,
+        314
+      ],
+      "flags": {},
+      "order": 8,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "image",
+          "name": "image",
+          "type": "COMBO",
+          "widget": {
+            "name": "image"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "choose file to upload",
+          "name": "upload",
+          "type": "IMAGEUPLOAD",
+          "widget": {
+            "name": "upload"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            4
+          ]
+        },
+        {
+          "localized_name": "MASK",
+          "name": "MASK",
+          "type": "MASK",
+          "slot_index": 1,
+          "links": []
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "LoadImage"
+      },
+      "widgets_values": [
+        "cmp_source.png",
+        "image"
+      ],
+      "title": "reference PHOTO"
+    },
+    {
+      "id": 9,
+      "type": "AnimaPoseControl",
+      "pos": [
+        680,
+        120
+      ],
+      "size": [
+        260,
+        200
+      ],
+      "flags": {},
+      "order": 9,
+      "mode": 0,
+      "properties": {
+        "Node name for S&R": "AnimaPoseControl"
+      },
+      "widgets_values": [
+        "R0_thin",
+        true,
+        true,
+        true,
+        true,
+        768,
+        ""
+      ],
+      "inputs": [
+        {
+          "name": "image",
+          "type": "IMAGE",
+          "link": 4
+        }
+      ],
+      "outputs": [
+        {
+          "name": "skeleton",
+          "type": "IMAGE",
+          "links": [
+            5,
+            26
+          ],
+          "slot_index": 0
+        },
+        {
+          "name": "keypoints",
+          "type": "POSE_KEYPOINT",
+          "links": [],
+          "slot_index": 1
+        }
+      ],
+      "title": "POSE: detect + style",
+      "color": "#533",
+      "bgcolor": "#223"
+    },
+    {
+      "id": 10,
+      "type": "ImageScale",
+      "pos": [
+        980,
+        120
+      ],
+      "size": [
+        300,
+        130
+      ],
+      "flags": {},
+      "order": 10,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "image",
+          "name": "image",
+          "type": "IMAGE",
+          "link": 5
+        },
+        {
+          "localized_name": "upscale_method",
+          "name": "upscale_method",
+          "type": "COMBO",
+          "widget": {
+            "name": "upscale_method"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "width",
+          "name": "width",
+          "type": "INT",
+          "widget": {
+            "name": "width"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "height",
+          "name": "height",
+          "type": "INT",
+          "widget": {
+            "name": "height"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "crop",
+          "name": "crop",
+          "type": "COMBO",
+          "widget": {
+            "name": "crop"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            6
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "ImageScale"
+      },
+      "widgets_values": [
+        "nearest-exact",
+        768,
+        768,
+        "disabled"
+      ]
+    },
+    {
+      "id": 11,
+      "type": "VAEEncode",
+      "pos": [
+        1280,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 11,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "pixels",
+          "name": "pixels",
+          "type": "IMAGE",
+          "link": 6
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 7
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            9
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEEncode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 12,
+      "type": "AnimaControlApply",
+      "pos": [
+        1580,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 12,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 8
+        },
+        {
+          "localized_name": "control_latent",
+          "name": "control_latent",
+          "type": "LATENT",
+          "link": 9
+        },
+        {
+          "localized_name": "control_embedder_path",
+          "name": "control_embedder_path",
+          "type": "STRING",
+          "widget": {
+            "name": "control_embedder_path"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "strength",
+          "name": "strength",
+          "type": "FLOAT",
+          "widget": {
+            "name": "strength"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            10
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "AnimaControlApply"
+      },
+      "widgets_values": [
+        "anima_pose_preview2.safetensors",
+        1
+      ]
+    },
+    {
+      "id": 13,
+      "type": "CFGZeroStar",
+      "pos": [
+        1880,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 13,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 10
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "patched_model",
+          "name": "patched_model",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            11
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CFGZeroStar"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 14,
+      "type": "KSampler",
+      "pos": [
+        2180,
+        120
+      ],
+      "size": [
+        300,
+        262
+      ],
+      "flags": {},
+      "order": 14,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 11
+        },
+        {
+          "localized_name": "positive",
+          "name": "positive",
+          "type": "CONDITIONING",
+          "link": 12
+        },
+        {
+          "localized_name": "negative",
+          "name": "negative",
+          "type": "CONDITIONING",
+          "link": 13
+        },
+        {
+          "localized_name": "latent_image",
+          "name": "latent_image",
+          "type": "LATENT",
+          "link": 14
+        },
+        {
+          "localized_name": "seed",
+          "name": "seed",
+          "type": "INT",
+          "widget": {
+            "name": "seed"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "steps",
+          "name": "steps",
+          "type": "INT",
+          "widget": {
+            "name": "steps"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "cfg",
+          "name": "cfg",
+          "type": "FLOAT",
+          "widget": {
+            "name": "cfg"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "sampler_name",
+          "name": "sampler_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "sampler_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "scheduler",
+          "name": "scheduler",
+          "type": "COMBO",
+          "widget": {
+            "name": "scheduler"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "denoise",
+          "name": "denoise",
+          "type": "FLOAT",
+          "widget": {
+            "name": "denoise"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            15
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "KSampler"
+      },
+      "widgets_values": [
+        2024,
+        "fixed",
+        30,
+        4,
+        "er_sde",
+        "simple",
+        1
+      ]
+    },
+    {
+      "id": 15,
+      "type": "VAEDecode",
+      "pos": [
+        2480,
+        120
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 15,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "samples",
+          "name": "samples",
+          "type": "LATENT",
+          "link": 15
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 16
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            17
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEDecode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 16,
+      "type": "SaveImage",
+      "pos": [
+        2780,
+        120
+      ],
+      "size": [
+        300,
+        270
+      ],
+      "flags": {},
+      "order": 16,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "images",
+          "name": "images",
+          "type": "IMAGE",
+          "link": 17
+        },
+        {
+          "localized_name": "filename_prefix",
+          "name": "filename_prefix",
+          "type": "STRING",
+          "widget": {
+            "name": "filename_prefix"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [],
+      "properties": {
+        "Node name for S&R": "SaveImage"
+      },
+      "widgets_values": [
+        "anima_pose_control"
+      ],
+      "title": "GENERATION"
+    },
+    {
+      "id": 17,
+      "type": "CFGZeroStar",
+      "pos": [
+        1880,
+        580
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 17,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 18
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "patched_model",
+          "name": "patched_model",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            19
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CFGZeroStar"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 18,
+      "type": "KSampler",
+      "pos": [
+        2180,
+        580
+      ],
+      "size": [
+        300,
+        262
+      ],
+      "flags": {},
+      "order": 18,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 19
+        },
+        {
+          "localized_name": "positive",
+          "name": "positive",
+          "type": "CONDITIONING",
+          "link": 20
+        },
+        {
+          "localized_name": "negative",
+          "name": "negative",
+          "type": "CONDITIONING",
+          "link": 21
+        },
+        {
+          "localized_name": "latent_image",
+          "name": "latent_image",
+          "type": "LATENT",
+          "link": 22
+        },
+        {
+          "localized_name": "seed",
+          "name": "seed",
+          "type": "INT",
+          "widget": {
+            "name": "seed"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "steps",
+          "name": "steps",
+          "type": "INT",
+          "widget": {
+            "name": "steps"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "cfg",
+          "name": "cfg",
+          "type": "FLOAT",
+          "widget": {
+            "name": "cfg"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "sampler_name",
+          "name": "sampler_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "sampler_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "scheduler",
+          "name": "scheduler",
+          "type": "COMBO",
+          "widget": {
+            "name": "scheduler"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "denoise",
+          "name": "denoise",
+          "type": "FLOAT",
+          "widget": {
+            "name": "denoise"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            23
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "KSampler"
+      },
+      "widgets_values": [
+        2024,
+        "fixed",
+        30,
+        4,
+        "er_sde",
+        "simple",
+        1
+      ]
+    },
+    {
+      "id": 19,
+      "type": "VAEDecode",
+      "pos": [
+        2480,
+        580
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 19,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "samples",
+          "name": "samples",
+          "type": "LATENT",
+          "link": 23
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 24
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            25
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEDecode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 20,
+      "type": "SaveImage",
+      "pos": [
+        2780,
+        580
+      ],
+      "size": [
+        300,
+        270
+      ],
+      "flags": {},
+      "order": 20,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "images",
+          "name": "images",
+          "type": "IMAGE",
+          "link": 25
+        },
+        {
+          "localized_name": "filename_prefix",
+          "name": "filename_prefix",
+          "type": "STRING",
+          "widget": {
+            "name": "filename_prefix"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [],
+      "properties": {
+        "Node name for S&R": "SaveImage"
+      },
+      "widgets_values": [
+        "anima_pose_BASE"
+      ],
+      "title": "BASE (no control)"
+    },
+    {
+      "id": 21,
+      "type": "SaveImage",
+      "pos": [
+        980,
+        580
+      ],
+      "size": [
+        300,
+        270
+      ],
+      "flags": {},
+      "order": 21,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "images",
+          "name": "images",
+          "type": "IMAGE",
+          "link": 26
+        },
+        {
+          "localized_name": "filename_prefix",
+          "name": "filename_prefix",
+          "type": "STRING",
+          "widget": {
+            "name": "filename_prefix"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [],
+      "properties": {
+        "Node name for S&R": "SaveImage"
+      },
+      "widgets_values": [
+        "anima_pose_SKELETON"
+      ],
+      "title": "SKELETON (control image)"
+    }
+  ],
+  "links": [
+    [
+      1,
+      1,
+      0,
+      2,
+      0,
+      "MODEL"
+    ],
+    [
+      2,
+      3,
+      0,
+      5,
+      0,
+      "CLIP"
+    ],
+    [
+      3,
+      3,
+      0,
+      6,
+      0,
+      "CLIP"
+    ],
+    [
+      4,
+      8,
+      0,
+      9,
+      0,
+      "IMAGE"
+    ],
+    [
+      5,
+      9,
+      0,
+      10,
+      0,
+      "IMAGE"
+    ],
+    [
+      6,
+      10,
+      0,
+      11,
+      0,
+      "IMAGE"
+    ],
+    [
+      7,
+      4,
+      0,
+      11,
+      1,
+      "VAE"
+    ],
+    [
+      8,
+      2,
+      0,
+      12,
+      0,
+      "MODEL"
+    ],
+    [
+      9,
+      11,
+      0,
+      12,
+      1,
+      "LATENT"
+    ],
+    [
+      10,
+      12,
+      0,
+      13,
+      0,
+      "MODEL"
+    ],
+    [
+      11,
+      13,
+      0,
+      14,
+      0,
+      "MODEL"
+    ],
+    [
+      12,
+      5,
+      0,
+      14,
+      1,
+      "CONDITIONING"
+    ],
+    [
+      13,
+      6,
+      0,
+      14,
+      2,
+      "CONDITIONING"
+    ],
+    [
+      14,
+      7,
+      0,
+      14,
+      3,
+      "LATENT"
+    ],
+    [
+      15,
+      14,
+      0,
+      15,
+      0,
+      "LATENT"
+    ],
+    [
+      16,
+      4,
+      0,
+      15,
+      1,
+      "VAE"
+    ],
+    [
+      17,
+      15,
+      0,
+      16,
+      0,
+      "IMAGE"
+    ],
+    [
+      18,
+      1,
+      0,
+      17,
+      0,
+      "MODEL"
+    ],
+    [
+      19,
+      17,
+      0,
+      18,
+      0,
+      "MODEL"
+    ],
+    [
+      20,
+      5,
+      0,
+      18,
+      1,
+      "CONDITIONING"
+    ],
+    [
+      21,
+      6,
+      0,
+      18,
+      2,
+      "CONDITIONING"
+    ],
+    [
+      22,
+      7,
+      0,
+      18,
+      3,
+      "LATENT"
+    ],
+    [
+      23,
+      18,
+      0,
+      19,
+      0,
+      "LATENT"
+    ],
+    [
+      24,
+      4,
+      0,
+      19,
+      1,
+      "VAE"
+    ],
+    [
+      25,
+      19,
+      0,
+      20,
+      0,
+      "IMAGE"
+    ],
+    [
+      26,
+      9,
+      0,
+      21,
+      0,
+      "IMAGE"
+    ]
+  ],
+  "groups": [
+    {
+      "id": 1,
+      "title": "CONTROL (pose LoRA)",
+      "bounding": [
+        1560,
+        60,
+        1260,
+        300
+      ],
+      "color": "#2a3f2a",
+      "font_size": 22,
+      "flags": {}
+    },
+    {
+      "id": 2,
+      "title": "BASE (no control) \u2014 same seed/prompt",
+      "bounding": [
+        1560,
+        520,
+        1260,
+        300
+      ],
+      "color": "#3f2a2a",
+      "font_size": 22,
+      "flags": {}
+    }
+  ],
+  "config": {},
+  "extra": {
+    "ds": {
+      "scale": 1.2284597357367357,
+      "offset": [
+        -896.6673282988922,
+        -2098.6519356376084
+      ]
+    }
+  },
+  "version": 0.4
+}

pose_control_edit.json

@@ -0,0 +1,1082 @@
+{
+  "id": "pose_control_edit",
+  "revision": 0,
+  "last_node_id": 16,
+  "last_link_id": 17,
+  "nodes": [
+    {
+      "id": 1,
+      "type": "UNETLoader",
+      "pos": [
+        20,
+        20
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 1,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "unet_name",
+          "name": "unet_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "unet_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "weight_dtype",
+          "name": "weight_dtype",
+          "type": "COMBO",
+          "widget": {
+            "name": "weight_dtype"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            1
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "UNETLoader"
+      },
+      "widgets_values": [
+        "anima-base-v1.0.safetensors",
+        "default"
+      ]
+    },
+    {
+      "id": 2,
+      "type": "CLIPLoader",
+      "pos": [
+        20,
+        160
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 2,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip_name",
+          "name": "clip_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "clip_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "type",
+          "name": "type",
+          "type": "COMBO",
+          "widget": {
+            "name": "type"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "device",
+          "name": "device",
+          "shape": 7,
+          "type": "COMBO",
+          "widget": {
+            "name": "device"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CLIP",
+          "name": "CLIP",
+          "type": "CLIP",
+          "slot_index": 0,
+          "links": [
+            2,
+            3
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPLoader"
+      },
+      "widgets_values": [
+        "qwen_3_06b_base.safetensors",
+        "stable_diffusion",
+        "default"
+      ]
+    },
+    {
+      "id": 3,
+      "type": "VAELoader",
+      "pos": [
+        20,
+        300
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 3,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "vae_name",
+          "name": "vae_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "vae_name"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "VAE",
+          "name": "VAE",
+          "type": "VAE",
+          "slot_index": 0,
+          "links": [
+            7,
+            16
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAELoader"
+      },
+      "widgets_values": [
+        "qwen_image_vae.safetensors"
+      ]
+    },
+    {
+      "id": 4,
+      "type": "LoraLoaderModelOnly",
+      "pos": [
+        20,
+        410
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 4,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 1
+        },
+        {
+          "localized_name": "lora_name",
+          "name": "lora_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "lora_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "strength_model",
+          "name": "strength_model",
+          "type": "FLOAT",
+          "widget": {
+            "name": "strength_model"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            8
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "LoraLoaderModelOnly"
+      },
+      "widgets_values": [
+        "anima_pose_preview2.safetensors",
+        1
+      ]
+    },
+    {
+      "id": 5,
+      "type": "CLIPTextEncode",
+      "pos": [
+        20,
+        540
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 5,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip",
+          "name": "clip",
+          "type": "CLIP",
+          "link": 2
+        },
+        {
+          "localized_name": "text",
+          "name": "text",
+          "type": "STRING",
+          "widget": {
+            "name": "text"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CONDITIONING",
+          "name": "CONDITIONING",
+          "type": "CONDITIONING",
+          "slot_index": 0,
+          "links": [
+            12
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPTextEncode"
+      },
+      "widgets_values": [
+        "newest, masterpiece, best quality, score_9, safe\n\n1girl, solo, simple background"
+      ]
+    },
+    {
+      "id": 6,
+      "type": "CLIPTextEncode",
+      "pos": [
+        20,
+        740
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 6,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "clip",
+          "name": "clip",
+          "type": "CLIP",
+          "link": 3
+        },
+        {
+          "localized_name": "text",
+          "name": "text",
+          "type": "STRING",
+          "widget": {
+            "name": "text"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "CONDITIONING",
+          "name": "CONDITIONING",
+          "type": "CONDITIONING",
+          "slot_index": 0,
+          "links": [
+            13
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CLIPTextEncode"
+      },
+      "widgets_values": [
+        "worst quality, low quality, score_1, score_2, blurry, jpeg artifacts"
+      ]
+    },
+    {
+      "id": 7,
+      "type": "EmptyLatentImage",
+      "pos": [
+        20,
+        940
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 7,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "width",
+          "name": "width",
+          "type": "INT",
+          "widget": {
+            "name": "width"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "height",
+          "name": "height",
+          "type": "INT",
+          "widget": {
+            "name": "height"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "batch_size",
+          "name": "batch_size",
+          "type": "INT",
+          "widget": {
+            "name": "batch_size"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            14
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "EmptyLatentImage"
+      },
+      "widgets_values": [
+        768,
+        768,
+        1
+      ]
+    },
+    {
+      "id": 8,
+      "type": "LoadImage",
+      "pos": [
+        360,
+        160
+      ],
+      "size": [
+        300,
+        314
+      ],
+      "flags": {},
+      "order": 8,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "image",
+          "name": "image",
+          "type": "COMBO",
+          "widget": {
+            "name": "image"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "choose file to upload",
+          "name": "upload",
+          "type": "IMAGEUPLOAD",
+          "widget": {
+            "name": "upload"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            4
+          ]
+        },
+        {
+          "localized_name": "MASK",
+          "name": "MASK",
+          "type": "MASK",
+          "slot_index": 1,
+          "links": []
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "LoadImage"
+      },
+      "widgets_values": [
+        "cmp_source.png",
+        "image"
+      ],
+      "title": "reference PHOTO"
+    },
+    {
+      "id": 9,
+      "type": "AnimaPoseControl",
+      "pos": [
+        680,
+        160
+      ],
+      "size": [
+        250,
+        180
+      ],
+      "flags": {},
+      "order": 9,
+      "mode": 0,
+      "properties": {
+        "Node name for S&R": "AnimaPoseControl"
+      },
+      "widgets_values": [
+        "R0_thin",
+        true,
+        true,
+        true,
+        true,
+        768,
+        ""
+      ],
+      "inputs": [
+        {
+          "name": "image",
+          "type": "IMAGE",
+          "link": 4
+        }
+      ],
+      "outputs": [
+        {
+          "name": "skeleton",
+          "type": "IMAGE",
+          "links": [
+            5
+          ],
+          "slot_index": 0
+        },
+        {
+          "name": "keypoints",
+          "type": "POSE_KEYPOINT",
+          "links": [],
+          "slot_index": 1
+        }
+      ],
+      "title": "POSE: detect + style",
+      "color": "#533",
+      "bgcolor": "#223"
+    },
+    {
+      "id": 10,
+      "type": "ImageScale",
+      "pos": [
+        1010,
+        160
+      ],
+      "size": [
+        300,
+        130
+      ],
+      "flags": {},
+      "order": 10,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "image",
+          "name": "image",
+          "type": "IMAGE",
+          "link": 5
+        },
+        {
+          "localized_name": "upscale_method",
+          "name": "upscale_method",
+          "type": "COMBO",
+          "widget": {
+            "name": "upscale_method"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "width",
+          "name": "width",
+          "type": "INT",
+          "widget": {
+            "name": "width"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "height",
+          "name": "height",
+          "type": "INT",
+          "widget": {
+            "name": "height"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "crop",
+          "name": "crop",
+          "type": "COMBO",
+          "widget": {
+            "name": "crop"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            6
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "ImageScale"
+      },
+      "widgets_values": [
+        "nearest-exact",
+        768,
+        768,
+        "disabled"
+      ]
+    },
+    {
+      "id": 11,
+      "type": "VAEEncode",
+      "pos": [
+        1210,
+        160
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 11,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "pixels",
+          "name": "pixels",
+          "type": "IMAGE",
+          "link": 6
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 7
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            9
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEEncode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 12,
+      "type": "AnimaControlApply",
+      "pos": [
+        1390,
+        160
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 12,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 8
+        },
+        {
+          "localized_name": "control_latent",
+          "name": "control_latent",
+          "type": "LATENT",
+          "link": 9
+        },
+        {
+          "localized_name": "control_embedder_path",
+          "name": "control_embedder_path",
+          "type": "STRING",
+          "widget": {
+            "name": "control_embedder_path"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "strength",
+          "name": "strength",
+          "type": "FLOAT",
+          "widget": {
+            "name": "strength"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "MODEL",
+          "name": "MODEL",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            10
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "AnimaControlApply"
+      },
+      "widgets_values": [
+        "anima_pose_preview2.safetensors",
+        1
+      ]
+    },
+    {
+      "id": 13,
+      "type": "CFGZeroStar",
+      "pos": [
+        1630,
+        160
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 13,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 10
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "patched_model",
+          "name": "patched_model",
+          "type": "MODEL",
+          "slot_index": 0,
+          "links": [
+            11
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "CFGZeroStar"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 14,
+      "type": "KSampler",
+      "pos": [
+        1790,
+        160
+      ],
+      "size": [
+        300,
+        262
+      ],
+      "flags": {},
+      "order": 14,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "model",
+          "name": "model",
+          "type": "MODEL",
+          "link": 11
+        },
+        {
+          "localized_name": "positive",
+          "name": "positive",
+          "type": "CONDITIONING",
+          "link": 12
+        },
+        {
+          "localized_name": "negative",
+          "name": "negative",
+          "type": "CONDITIONING",
+          "link": 13
+        },
+        {
+          "localized_name": "latent_image",
+          "name": "latent_image",
+          "type": "LATENT",
+          "link": 14
+        },
+        {
+          "localized_name": "seed",
+          "name": "seed",
+          "type": "INT",
+          "widget": {
+            "name": "seed"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "steps",
+          "name": "steps",
+          "type": "INT",
+          "widget": {
+            "name": "steps"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "cfg",
+          "name": "cfg",
+          "type": "FLOAT",
+          "widget": {
+            "name": "cfg"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "sampler_name",
+          "name": "sampler_name",
+          "type": "COMBO",
+          "widget": {
+            "name": "sampler_name"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "scheduler",
+          "name": "scheduler",
+          "type": "COMBO",
+          "widget": {
+            "name": "scheduler"
+          },
+          "link": null
+        },
+        {
+          "localized_name": "denoise",
+          "name": "denoise",
+          "type": "FLOAT",
+          "widget": {
+            "name": "denoise"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "LATENT",
+          "name": "LATENT",
+          "type": "LATENT",
+          "slot_index": 0,
+          "links": [
+            15
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "KSampler"
+      },
+      "widgets_values": [
+        2024,
+        "fixed",
+        30,
+        4,
+        "er_sde",
+        "simple",
+        1
+      ]
+    },
+    {
+      "id": 15,
+      "type": "VAEDecode",
+      "pos": [
+        2050,
+        160
+      ],
+      "size": [
+        300,
+        120
+      ],
+      "flags": {},
+      "order": 15,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "samples",
+          "name": "samples",
+          "type": "LATENT",
+          "link": 15
+        },
+        {
+          "localized_name": "vae",
+          "name": "vae",
+          "type": "VAE",
+          "link": 16
+        }
+      ],
+      "outputs": [
+        {
+          "localized_name": "IMAGE",
+          "name": "IMAGE",
+          "type": "IMAGE",
+          "slot_index": 0,
+          "links": [
+            17
+          ]
+        }
+      ],
+      "properties": {
+        "Node name for S&R": "VAEDecode"
+      },
+      "widgets_values": []
+    },
+    {
+      "id": 16,
+      "type": "SaveImage",
+      "pos": [
+        2230,
+        160
+      ],
+      "size": [
+        300,
+        270
+      ],
+      "flags": {},
+      "order": 16,
+      "mode": 0,
+      "inputs": [
+        {
+          "localized_name": "images",
+          "name": "images",
+          "type": "IMAGE",
+          "link": 17
+        },
+        {
+          "localized_name": "filename_prefix",
+          "name": "filename_prefix",
+          "type": "STRING",
+          "widget": {
+            "name": "filename_prefix"
+          },
+          "link": null
+        }
+      ],
+      "outputs": [],
+      "properties": {
+        "Node name for S&R": "SaveImage"
+      },
+      "widgets_values": [
+        "pose_edit_out"
+      ],
+      "title": "GENERATION (v2)"
+    }
+  ],
+  "links": [
+    [
+      1,
+      1,
+      0,
+      4,
+      0,
+      "MODEL"
+    ],
+    [
+      2,
+      2,
+      0,
+      5,
+      0,
+      "CLIP"
+    ],
+    [
+      3,
+      2,
+      0,
+      6,
+      0,
+      "CLIP"
+    ],
+    [
+      4,
+      8,
+      0,
+      9,
+      0,
+      "IMAGE"
+    ],
+    [
+      5,
+      9,
+      0,
+      10,
+      0,
+      "IMAGE"
+    ],
+    [
+      6,
+      10,
+      0,
+      11,
+      0,
+      "IMAGE"
+    ],
+    [
+      7,
+      3,
+      0,
+      11,
+      1,
+      "VAE"
+    ],
+    [
+      8,
+      4,
+      0,
+      12,
+      0,
+      "MODEL"
+    ],
+    [
+      9,
+      11,
+      0,
+      12,
+      1,
+      "LATENT"
+    ],
+    [
+      10,
+      12,
+      0,
+      13,
+      0,
+      "MODEL"
+    ],
+    [
+      11,
+      13,
+      0,
+      14,
+      0,
+      "MODEL"
+    ],
+    [
+      12,
+      5,
+      0,
+      14,
+      1,
+      "CONDITIONING"
+    ],
+    [
+      13,
+      6,
+      0,
+      14,
+      2,
+      "CONDITIONING"
+    ],
+    [
+      14,
+      7,
+      0,
+      14,
+      3,
+      "LATENT"
+    ],
+    [
+      15,
+      14,
+      0,
+      15,
+      0,
+      "LATENT"
+    ],
+    [
+      16,
+      3,
+      0,
+      15,
+      1,
+      "VAE"
+    ],
+    [
+      17,
+      15,
+      0,
+      16,
+      0,
+      "IMAGE"
+    ]
+  ],
+  "groups": [],
+  "config": {},
+  "extra": {
+    "ds": {
+      "scale": 1.2284597357367357,
+      "offset": [
+        -896.6673282988922,
+        -2098.6519356376084
+      ]
+    }
+  },
+  "version": 0.4
+}

scripts/render_skeletons.py

@@ -0,0 +1,44 @@
+"""Render full-body DWPose skeletons (COCO-WholeBody 133-kpt, black background) as control images.
+Reuses the same DWPose inference as foot_adetailer/scripts/autolabel_dwpose.py.
+NOTE: rtmlib's draw_skeleton only supports openpose_skeleton=True for 17/18/134/26 kpts; the
+Wholebody 133-kpt output must be drawn with openpose_skeleton=False (train + infer on the same style)."""
+import argparse, glob, os
+import numpy as np
+
+
+def render_one(img, pose_fn, draw_fn, out_path, kp_thr=0.3):
+    import cv2    # lazy import so the module stays importable without opencv installed
+    keypoints, scores = pose_fn(img)
+    canvas = np.zeros_like(img)
+    canvas = draw_fn(canvas, keypoints, scores, openpose_skeleton=False, kpt_thr=kp_thr)
+    cv2.imwrite(str(out_path), canvas)
+
+
+def main():
+    ap = argparse.ArgumentParser()
+    ap.add_argument('--imgs', required=True)
+    ap.add_argument('--out', required=True)
+    ap.add_argument('--kp-thr', type=float, default=0.3)
+    ap.add_argument('--device', default='cuda')
+    ap.add_argument('--limit', type=int, default=0)
+    args = ap.parse_args()
+
+    import cv2
+    from rtmlib import Wholebody, draw_skeleton    # lazy: not in local .venv
+    pose = Wholebody(to_openpose=False, mode='balanced', backend='onnxruntime', device=args.device)
+    os.makedirs(args.out, exist_ok=True)
+    files = sorted(sum((glob.glob(os.path.join(args.imgs, f'*.{e}')) for e in ('jpg', 'jpeg', 'png', 'webp')), []))
+    if args.limit:
+        files = files[:args.limit]
+    for i, fp in enumerate(files):
+        img = cv2.imread(fp)
+        if img is None:
+            continue
+        stem = os.path.splitext(os.path.basename(fp))[0]
+        render_one(img, lambda im: pose(im), draw_skeleton, os.path.join(args.out, stem + '.png'), args.kp_thr)
+        if i % 200 == 0:
+            print(f'{i}/{len(files)}')
+
+
+if __name__ == '__main__':
+    main()