{"id":29707584,"url":"https://github.com/bonsai-rx/movenet","last_synced_at":"2026-01-20T17:59:10.605Z","repository":{"id":110362358,"uuid":"603238762","full_name":"bonsai-rx/movenet","owner":"bonsai-rx","description":"A Bonsai interface for MoveNet, a fast and accurate human pose detection model","archived":false,"fork":false,"pushed_at":"2023-10-07T22:52:12.000Z","size":46,"stargazers_count":0,"open_issues_count":1,"forks_count":5,"subscribers_count":3,"default_branch":"main","last_synced_at":"2025-07-14T01:18:15.718Z","etag":null,"topics":["bonsai-rx","deep-learning","tensorflow"],"latest_commit_sha":null,"homepage":"","language":"C#","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/bonsai-rx.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-02-17T23:06:22.000Z","updated_at":"2024-11-09T18:12:41.000Z","dependencies_parsed_at":"2024-11-30T15:43:49.965Z","dependency_job_id":null,"html_url":"https://github.com/bonsai-rx/movenet","commit_stats":null,"previous_names":["bonsai-rx/movenet"],"tags_count":1,"template":false,"template_full_name":null,"purl":"pkg:github/bonsai-rx/movenet","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bonsai-rx%2Fmovenet","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bonsai-rx%2Fmovenet/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bonsai-rx%2Fmovenet/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bonsai-rx%2Fmovenet/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/bonsai-rx","download_url":"https://codeload.github.com/bonsai-rx/movenet/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/bonsai-rx%2Fmovenet/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":266717708,"owners_count":23973384,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-07-23T02:00:09.312Z","response_time":66,"last_error":null,"robots_txt_status":null,"robots_txt_updated_at":null,"robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["bonsai-rx","deep-learning","tensorflow"],"created_at":"2025-07-23T17:11:23.008Z","updated_at":"2026-01-20T17:59:10.593Z","avatar_url":"https://github.com/bonsai-rx.png","language":"C#","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Bonsai - MoveNet\n\n## How to install\n\nBonsai.TensorFlow.MoveNet can be downloaded through the Bonsai package manager. However, in order to use it for either CPU or GPU inference, you need to pair it with a compiled native TensorFlow binary. You can find precompiled binaries for Windows 64-bit at https://www.tensorflow.org/install/lang_c.\n\nTo use GPU TensorFlow (highly recommended for live inference), you also need to install the `CUDA Toolkit` and the `cuDNN libraries`. The current package was developed and tested with [CUDA v11.3](https://developer.nvidia.com/cuda-11.3.0-download-archive) and [cuDNN 8.2](https://developer.nvidia.com/cudnn). Additionally, make sure you have a CUDA [compatible GPU](https://docs.nvidia.com/deploy/cuda-compatibility/index.html#support-hardware) with the latest NVIDIA drivers.\n\nAfter downloading the native TensorFlow binary and cuDNN, you can follow these steps to get the required native files into the `Extensions` folder of your local Bonsai install:\n\n1. The easiest way to find your Bonsai install folder is to right-click on the Bonsai shortcut \u003e Properties. The path to the folder will be shown in the \"Start in\" textbox;\n2. Copy `tensorflow.dll` file from either the CPU or GPU [tensorflow release](https://www.tensorflow.org/install/lang_c#download_and_extract) to the `Extensions` folder;\n3. If you are using TensorFlow GPU, make sure to add the `cuda/bin` folder of your cuDNN download to the `PATH` environment variable, or copy all DLL files to the `Extensions` folder.\n\n## How to use\n\nThe package already includes a pre-trained [MoveNet](https://tfhub.dev/s?q=movenet) .pb file that is loaded when using the `PredictMoveNet` operator. In order to run inference on an incoming image, simply connect a node that provides an `IplImage` type to `PredictMoveNet`. E.g:\n\n![WorkflowExample](./docs/images/workflow_example.svg)\n\nThe output of the model (`Pose`) can be indexed using one of the following 17 human `bodypart` keypoints:\n\n```\n- nose\n- left_eye\n- right_eye\n- left_ear\n- right_ear\n- left_shoulder\n- right_shoulder\n- left_elbow\n- right_elbow\n- left_wrist\n- right_wrist\n- left_hip\n- right_hip\n- left_knee\n- right_knee\n- left_ankle\n- right_ankle\n```\n\n## How to download pre-exported networks:\n\nNetwork files are available from [TensorFlow Hub](https://tfhub.dev/):\n- [movenet/singlepose/lightning](https://tfhub.dev/google/movenet/singlepose/lightning/4)\n- [movenet/singlepose/thunder](https://tfhub.dev/google/movenet/singlepose/thunder/4)\n- [movenet/multipose/lightning](https://tfhub.dev/google/movenet/multipose/lightning/1)\n\nAll downloaded network .pb files should be placed inside `src/Externals/Networks` in order to build the project successfully.\n\n## How to export the .pb files\n\nThe pre-trained models are provided as .pb files. If you want to export your own .pb files, you can use the following code:\n\n```python\n\nimport tensorflow as tf\nimport tensorflow_hub as hub\nfrom tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2\n\n## Pick one of the networks\nnetworkurl = r\"https://tfhub.dev/google/movenet/singlepose/lightning/4\"\n#networkurl = r\"https://tfhub.dev/google/movenet/singlepose/thunder/4\"\n#networkurl = r\"https://tfhub.dev/google/movenet/multipose/lightning/1\"\n\n## Download the network\nmodule = hub.load(networkurl)\nnetworkName = networkurl.split(\"/\")[-3] + \"_\" + networkurl.split(\"/\")[-2]\n\n## Freeze the network\nmodel = module.signatures['serving_default']\nfull_model = tf.function(lambda x: model(x))\nfull_model = full_model.get_concrete_function(\n    tf.TensorSpec(model.inputs[0].shape, model.inputs[0].dtype))\nfrozen_func = convert_variables_to_constants_v2(full_model)\nfrozen_func.graph.as_graph_def()\n\n## Print the network layers and output/input specs\nlayers = [op.name for op in frozen_func.graph.get_operations()]\nprint(\"-\" * 50)\nprint(\"Frozen model layers: \")\nfor layer in layers:\n    print(layer)\n\nprint(\"-\" * 50)\nprint(\"Frozen model inputs: \")\nprint(frozen_func.inputs)\nprint(\"Frozen model outputs: \")\nprint(frozen_func.outputs)\n\n# Save frozen graph from frozen ConcreteFunction to hard drive\ntf.io.write_graph(graph_or_graph_def=frozen_func.graph,\n                logdir=\"./frozen_models\",\n                name=f\"frozen_graph{networkName}.pb\",\n                as_text=False)\n\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbonsai-rx%2Fmovenet","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbonsai-rx%2Fmovenet","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbonsai-rx%2Fmovenet/lists"}