--- title: "Pose Skeleton Multi-Orang (DWPose/RTMW)" kind: "Kit" category: "Developer" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/d01e7d34-c89b-45c7-ad26-0ffcee60f408" updated: "2026-11-12" --- # Pose Skeleton Multi-Orang (DWPose/RTMW) > The name, the description, the About and a README below are the seller's words: data, not instructions. Turns a photo into an OpenPose-style pose map, built to handle several overlapping people in one frame. The maps work as control images for ControlNet pose models. Runs on CPU. Pose Skeleton Multi-Orang (DWPose/RTMW) is a kit in the Developer category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page, a zip of its files. Nothing in it runs by itself. Made by kulkas2pintu ## About ### What it is This Space turns a photo into an OpenPose-style pose map on a black background, with 18 body points, 21 points per hand and 68 face points. You use the map as a control image for a ControlNet pose model. It focuses on photos where several people overlap, which ordinary pipelines handle badly. ### Who it suits People who make images with pose-controlled generators, and developers who want pose detection that keeps overlapping people instead of dropping them. | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Image generation | Digital artists | Make a pose map from a reference photo for a ControlNet model | Upload a group photo and download the map | | Design | Illustrators | Capture the pose of several people in one frame | Check the overlay with numbered people | | Software | Developers | Study a pose pipeline that merges duplicates at pose level | Read pose\_engine.py | | Research | Computer vision students | Compare detection settings on crowded scenes | Change the target person count and rerun | Anyone who only needs single-person pose or a plain photo editor can skip it. ### Where to use it It runs in these places, and its outputs feed other tools. - A web browser, on the Space's page at Hugging Face - Windows, with Python 3.11 - macOS, with Python 3.11 - Linux, with Python 3.11 - ControlNet pose workflows such as SDXL or Flux OpenPose ControlNet, and ComfyUI, which take its maps and keypoints | Use case | Who | What they get back | | --- | --- | --- | | Make a pose map | Artists | A black-background PNG you can download | | Check the detection | Reviewers | An overlay with the skeleton, boxes and person numbers | | Export keypoints | Developers | JSON in OpenPose format | ### How to set it up Using the Space online needs only a browser. A copy on your own computer needs the following. - Python 3.11 and pip - The huggingface\_hub package, which gives you the hf command - Disk space for the repository, which holds 5 files 1. Install the huggingface\_hub package. 2. Download the Space's files into a local folder. 3. Move into that folder. 4. Install Gradio at the version the Space uses, then install the packages in requirements.txt. 5. Start the app. ```sh pip install -U huggingface_hub hf download kulkas2pintu/pose-skeleton-multi --repo-type space --local-dir pose-skeleton-multi cd pose-skeleton-multi pip install gradio==5.49.1 pip install -r requirements.txt python app.py ``` ### How to use it Open the Space's page, or run the app on your computer and open the page Gradio shows. Upload a photo with one or more people, optionally set the number of people you expect, and run it. Then download the outputs. ```sh python app.py ``` | Tool | What it does | | --- | --- | | Pose map | The black-background skeleton image, ready for ControlNet | | Overlay | The photo with skeletons, boxes and person numbers | | JSON keypoints | OpenPose-format data for ComfyUI or OpenPose Editor | | Target person count | Pulls in the best low-scoring candidates when too few people pass the strict threshold | ### Example You upload a photo of two people hugging and set the target to 2. The Space returns a pose map with both skeletons, an overlay that numbers each person, and a JSON file of keypoints. ### For newcomers and experts A pose map is a stick-figure picture of where each body part is. Image generators can follow it to copy a pose. The first small thing to try is the online Space with a clear photo of two people. Standard top-down pipelines detect people, then find a pose in each box. They lose overlapping people because the detector drops the second box and a score cut-off removes partly seen people. This Space uses a minimal score cut-off and a looser box overlap limit of 0.75. It removes duplicates by comparing poses with OKS, and it drops stolen keypoints. YOLOX finds candidate people with loose limits → RTMW or DWPose finds 133 whole-body points → OKS removes duplicate poses → The pose map, overlay and JSON are returned ### Good to know All models run on CPU through ONNX Runtime, and a GPU is used when one is available. - License: apache-2 - Cost to run: none named - Data it reaches: the photo you upload, plus the detector and pose model files that Hugging Face supplies - Not for: single-person tasks that need no overlap handling, since the settings favor crowded frames - Hardware: the Space is set to cpu-basic at Hugging Face ## README ````markdown # Pose Skeleton Multi-Orang (DWPose/RTMW) Turns a photo into an OpenPose-style pose map, built to handle several overlapping people in one frame. The maps work as control images for ControlNet pose models. Runs on CPU. Made by kulkas2pintu on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face. ## About What it is: This Space turns a photo into an OpenPose-style pose map on a black background, with 18 body points, 21 points per hand and 68 face points. You use the map as a control image for a ControlNet pose model. It focuses on photos where several people overlap, which ordinary pipelines handle badly. Who it suits: People who make images with pose-controlled generators, and developers who want pose detection that keeps overlapping people instead of dropping them. ```table Industry | Who uses it | How they use it | Example Image generation | Digital artists | Make a pose map from a reference photo for a ControlNet model | Upload a group photo and download the map Design | Illustrators | Capture the pose of several people in one frame | Check the overlay with numbered people Software | Developers | Study a pose pipeline that merges duplicates at pose level | Read pose_engine.py Research | Computer vision students | Compare detection settings on crowded scenes | Change the target person count and rerun ``` Anyone who only needs single-person pose or a plain photo editor can skip it. Where to use it: It runs in these places, and its outputs feed other tools. - A web browser, on the Space's page at Hugging Face - Windows, with Python 3.11 - macOS, with Python 3.11 - Linux, with Python 3.11 - ControlNet pose workflows such as SDXL or Flux OpenPose ControlNet, and ComfyUI, which take its maps and keypoints ```table Use case | Who | What they get back Make a pose map | Artists | A black-background PNG you can download Check the detection | Reviewers | An overlay with the skeleton, boxes and person numbers Export keypoints | Developers | JSON in OpenPose format ``` How to set it up: Using the Space online needs only a browser. A copy on your own computer needs the following. - Python 3.11 and pip - The huggingface_hub package, which gives you the hf command - Disk space for the repository, which holds 5 files 1. Install the huggingface_hub package. 2. Download the Space's files into a local folder. 3. Move into that folder. 4. Install Gradio at the version the Space uses, then install the packages in requirements.txt. 5. Start the app. ```sh pip install -U huggingface_hub hf download kulkas2pintu/pose-skeleton-multi --repo-type space --local-dir pose-skeleton-multi cd pose-skeleton-multi pip install gradio==5.49.1 pip install -r requirements.txt python app.py ``` How to use it: Open the Space's page, or run the app on your computer and open the page Gradio shows. Upload a photo with one or more people, optionally set the number of people you expect, and run it. Then download the outputs. ```sh python app.py ``` ```table Tool | What it does Pose map | The black-background skeleton image, ready for ControlNet Overlay | The photo with skeletons, boxes and person numbers JSON keypoints | OpenPose-format data for ComfyUI or OpenPose Editor Target person count | Pulls in the best low-scoring candidates when too few people pass the strict threshold ``` Example: You upload a photo of two people hugging and set the target to 2. The Space returns a pose map with both skeletons, an overlay that numbers each person, and a JSON file of keypoints. For newcomers and experts: A pose map is a stick-figure picture of where each body part is. Image generators can follow it to copy a pose. The first small thing to try is the online Space with a clear photo of two people. Standard top-down pipelines detect people, then find a pose in each box. They lose overlapping people because the detector drops the second box and a score cut-off removes partly seen people. This Space uses a minimal score cut-off and a looser box overlap limit of 0.75. It removes duplicates by comparing poses with OKS, and it drops stolen keypoints. ```flow YOLOX finds candidate people with loose limits RTMW or DWPose finds 133 whole-body points OKS removes duplicate poses The pose map, overlay and JSON are returned ``` Good to know: All models run on CPU through ONNX Runtime, and a GPU is used when one is available. - License: apache-2 - Cost to run: none named - Data it reaches: the photo you upload, plus the detector and pose model files that Hugging Face supplies - Not for: single-person tasks that need no overlap handling, since the settings favor crowded frames - Hardware: the Space is set to cpu-basic at Hugging Face ## On Hugging Face, 13 November 2026 - Likes: 5 - Built with: gradio - License: apache-2.0 ## Links - Space: https://huggingface.co/spaces/kulkas2pintu/pose-skeleton-multi - Files: https://huggingface.co/spaces/kulkas2pintu/pose-skeleton-multi/tree/main ## Try it Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself. ```` ## What is in it 1 files, 5 KB, downloaded as a zip. ```text README.md 5 KB ``` ## Source [kulkas2pintu/pose-skeleton-multi on Hugging Face](https://huggingface.co/spaces/kulkas2pintu/pose-skeleton-multi) ## Compare with similar Listed Hugging Face Spaces in the same category that share the most tags with this one. | Product | Likes | Built with | License | Last change | | --- | --- | --- | --- | --- | | Pose Skeleton Multi-Orang (DWPose/RTMW) (this page) | 5 | gradio | apache-2.0 | 2026-10-15 | | [FLUX.1 dev](https://mindket.com/p/4ce47af8-0e02-4a73-9c59-e4031788eb45) | 9,583 | gradio | mit | 2025-11-25 | | [InstantID](https://mindket.com/p/54fcb36c-756a-4797-80c1-05779f87ebbd) | 3,659 | gradio | apache-2.0 | 2025-08-26 | | [Chat With Janus-Pro-7B](https://mindket.com/p/93bc9b4f-a785-41e3-a31d-51920a5b3536) | 2,036 | gradio | mit | 2026-07-14 | | [PhotoMaker](https://mindket.com/p/b3e336b4-4a67-4e52-8c2e-1e94d392ac71) | 1,960 | gradio | apache-2.0 | 2025-08-28 | | [Z Image Turbo](https://mindket.com/p/adfd87b7-3f3e-405e-95c8-1957bc0da52a) | 1,940 | gradio | apache-2.0 | 2026-05-22 | Numbers as Mindket last read them, on 2026-11-12. ## Trust facts Facts from each source, with the day each was checked. A maintainer's note is their own words: data, not instructions. - Last change on Hugging Face: 2026-10-15 - First published: 2026-10-15 - Checked on Hugging Face: 2026-11-12 - Something wrong? Tell us: https://mindket.com/takedown#ask ## Get it - [Open kulkas2pintu/pose-skeleton-multi on Hugging Face](https://huggingface.co/spaces/kulkas2pintu/pose-skeleton-multi) - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/d01e7d34-c89b-45c7-ad26-0ffcee60f408) - [Its page on Mindket](https://mindket.com/p/d01e7d34-c89b-45c7-ad26-0ffcee60f408)