[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7A79ME5REJ3M3BD5CSJ4DWD":3,"skill-same:01M7A79ME5REJ3M3BD5CSJ4DWD":98},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":32,"commitSha":33,"skillDirectory":34,"forksCount":35,"repositoryHtmlUrl":36,"repositoryHomepageUrl":37,"ownerHtmlUrl":38,"ownerBlogUrl":39,"ownerDescription":40,"updatedAt":41,"install":42,"resources":55},"01M7A79ME5REJ3M3BD5CSJ4DWD","tao-train-pose-classification",null,"Pose classification using ST-GCN (Spatial Temporal Graph Convolutional Network). Classifies skeleton sequences into action categories from pose-keypoint data. Use when training, evaluating, exporting, or running inference for a TAO pose-classification model. Trigger phrases include \"train pose classification\", \"skeleton action recognition\", \"ST-GCN\", \"keypoint sequence classifier\".","themusebook\u002Fskills","themusebook","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F349927247?v=4","01M7A5C1RV5S3V5FF6ZACVDNRN","musebook","01M7A5C1SFC3KRZAXNN60XJ7KM",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: tao-train-pose-classification\ndescription: Pose classification using ST-GCN (Spatial Temporal Graph Convolutional Network). Classifies skeleton sequences\n  into action categories from pose-keypoint data. Use when training, evaluating, exporting, or running inference for a TAO\n  pose-classification model. Trigger phrases include \"train pose classification\", \"skeleton action recognition\", \"ST-GCN\",\n  \"keypoint sequence classifier\".\nlicense: Apache-2.0\ncompatibility: Requires docker + nvidia-container-toolkit.\nmetadata:\n  version: \"0.1.0\"\n  author: NVIDIA Corporation\nallowed-tools: Read Bash\ntags:\n- pose\n- classification\n---\n\n# Pose Classification\n\nPose classification using ST-GCN (Spatial Temporal Graph Convolutional Network). Classifies skeleton sequences into action categories from pose keypoint data.\n\nTypically trained from scratch on skeleton data.\n\nThe packaged PyTorch Pose Classification CLI supports `dataset_convert`, `train`, `evaluate`, `export`, and `inference`. `dataset_convert` is conditional: run it only when the input is raw DeepStream BodyPose JSON. If the dataset is already converted to TAO-ready `.npy` \u002F `.pkl` files, start directly with `train` on those files and mark dataset conversion as `not run: preconverted dataset provided` in validation reports. This model does not expose deploy, prune, quantize, or standalone retrain actions. Resume\u002Fretrain behavior uses `pose_classification train -e ...` with `train.resume_training_checkpoint_path` populated.\n\n## Dataclass Schemas\n\nGenerated TAO Core schemas are packaged in `schemas\u002F\u003Caction>.schema.json`, with `schemas\u002Fmanifest.json` listing available actions. Each generated schema also emits `references\u002Fspec_template_\u003Caction>.yaml` from the schema top-level `default` field. AutoML enablement is declared at the model layer in `references\u002Fskill_info.yaml` via `automl_enabled`. Runnable AutoML still requires `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` to exist and parse. Use the packaged train schema for `automl_default_parameters`, `automl_disabled_parameters`, defaults, min\u002Fmax bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect `~\u002Ftao-core` at runtime; maintainers regenerate schemas\u002Ftemplates before packaging the skill bank.\n\n## Train Action Policy\n\nThis model is AutoML-enabled at the model layer. Before handling any train-stage request, read `references\u002Fskill_info.yaml` and resolve the run override from either an explicit `automl_policy` value or the user's workflow request. Use `automl_policy: on` by default and only expose `on` \u002F `off` in new launch prompts. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only. When `automl_policy: on`, `automl_enabled: true`, and both `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` are packaged, route the train action through `tao-skill-bank:tao-run-automl` by default with this model's `skill_dir`. Preserve workflow\u002Fapplication overrides for datasets, specs, output directories, GPU\u002Fplatform settings, parent checkpoints, and `automl_policy`. Use direct model training only when `automl_policy: off` or the packaged train schema\u002Ftemplate is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.\n\nNon-train actions such as `evaluate`, `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Training Requirements\n\n- **Dataset type:** pose_classification\n- **Formats:** default\n- **Monitoring metric:** val_loss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| dataset_convert (optional) | dataset_convert.data | id | DeepStream BodyPose JSON | No |\n| evaluate | evaluate.test_dataset.data_path | train_datasets | val_data.npy | No |\n| evaluate | evaluate.test_dataset.label_path | train_datasets | val_label.pkl | No |\n| inference | inference.test_dataset.data_path | train_datasets | test_data.npy | No |\n| train | dataset.train_dataset.data_path | train_datasets | train_data.npy | No |\n| train | dataset.train_dataset.label_path | train_datasets | train_label.pkl | No |\n| train | dataset.val_dataset.data_path | train_datasets | val_data.npy | No |\n| train | dataset.val_dataset.label_path | train_datasets | val_label.pkl | No |\n\n### Typical Spec Overrides\n\nData source overrides are **mandatory for every action being run** — the agent MUST construct data source paths from the Per-Action Dataset Requirements table above and include them in `spec_overrides`. Do not run `dataset_convert` when the supplied dataset is already converted to `.npy` \u002F `.pkl` files.\n\n```python\nS3_TRAIN = \"s3:\u002F\u002Fbucket\u002Fdata\u002Fpurpose_built_models_pose_classification_train\u002Fnvidia\"\nCHECKPOINT = \"\u002Fresults\u002F{train_job_id}\u002Fresults_dir\u002Fmodel_epoch_000_step_00007.pth\"\n```\n\n**dataset_convert (optional; raw DeepStream BodyPose JSON only):**\n```python\n{\n    \"dataset_convert.data\": \"s3:\u002F\u002Fbucket\u002Fdata\u002F\u003Cdeepstream-bodypose-output>.json\",\n}\n```\n\n**train (mandatory data sources):**\n```python\n{\n    \"train.num_epochs\": 30,\n    \"train.checkpoint_interval\": 10,\n    \"train.validation_interval\": 10,\n    \"train.num_gpus\": 1,\n    \"wandb.enable\": False,\n    \"dataset.num_classes\": 6,\n    \"dataset.label_map\": {\n        \"class_0\": 0,\n        \"class_1\": 1,\n        \"class_2\": 2,\n        \"class_3\": 3,\n        \"class_4\": 4,\n        \"class_5\": 5,\n    },\n    \"model.graph_layout\": \"nvidia\",\n    \"dataset.train_dataset.data_path\": f\"{S3_TRAIN}\u002Ftrain_data.npy\",\n    \"dataset.train_dataset.label_path\": f\"{S3_TRAIN}\u002Ftrain_label.pkl\",\n    \"dataset.val_dataset.data_path\": f\"{S3_TRAIN}\u002Fval_data.npy\",\n    \"dataset.val_dataset.label_path\": f\"{S3_TRAIN}\u002Fval_label.pkl\",\n}\n```\n\n**resume train (mandatory checkpoint):**\n```python\n{\n    \"train.num_epochs\": 31,\n    \"train.resume_training_checkpoint_path\": CHECKPOINT,\n    \"dataset.train_dataset.data_path\": f\"{S3_TRAIN}\u002Ftrain_data.npy\",\n    \"dataset.train_dataset.label_path\": f\"{S3_TRAIN}\u002Ftrain_label.pkl\",\n    \"dataset.val_dataset.data_path\": f\"{S3_TRAIN}\u002Fval_data.npy\",\n    \"dataset.val_dataset.label_path\": f\"{S3_TRAIN}\u002Fval_label.pkl\",\n}\n```\n\n**evaluate (mandatory data sources):**\n```python\n{\n    \"evaluate.test_dataset.data_path\": f\"{S3_TRAIN}\u002Fval_data.npy\",\n    \"evaluate.test_dataset.label_path\": f\"{S3_TRAIN}\u002Fval_label.pkl\",\n    \"evaluate.checkpoint\": CHECKPOINT,\n}\n```\n\n**export (mandatory checkpoint and output):**\n```python\n{\n    \"export.checkpoint\": CHECKPOINT,\n    \"export.onnx_file\": \"\u002Fresults\u002F{export_job_id}\u002Fresults_dir\u002Fpose_classification.onnx\",\n}\n```\n\n**inference (mandatory data sources):**\n```python\n{\n    \"inference.test_dataset.data_path\": f\"{S3_TRAIN}\u002Ftest_data.npy\",\n    \"inference.test_dataset.label_path\": f\"{S3_TRAIN}\u002Ftest_label.pkl\",\n    \"inference.checkpoint\": CHECKPOINT,\n    \"inference.output_file\": \"\u002Fresults\u002Fpose_classification_inference.txt\",\n}\n```\n## Dataset Convert\n\nDataset conversion is optional for Pose Classification. Run `pose_classification dataset_convert` only when the user supplies raw DeepStream BodyPose JSON. For the common S3 validation dataset, the data is already converted to `train_data.npy`, `train_label.pkl`, `val_data.npy`, `val_label.pkl`, `test_data.npy`, and `test_label.pkl`; use those files directly for train\u002Fevaluate\u002Finference\u002Fexport flows and do not synthesize fake BodyPose JSON.\n\n## Eval Dataset\n\nOptional. Validation data is provided alongside training as val_data.npy \u002F val_label.pkl. TAO training emits `val_loss` as the TensorBoard validation scalar for this model; use `val_loss` with minimize direction for AutoML selection unless a custom evaluation hook supplies a different metric.\n\n## Important Parameters\n\n- **dataset.num_classes**: Number of pose action classes. Default 6.\n- **model.graph_layout**: Skeleton graph layout. Options: nvidia, openpose. Determines joint connectivity.\n- **model.graph_strategy**: Graph partitioning strategy for GCN.\n- **train.optim.lr**: Learning rate. Default 0.1 (SGD). Higher than vision models due to graph convolution properties.\n- **model.dropout**: Dropout rate for regularization.\n\n## Multi-GPU \u002F Multi-Node\n\n**Launch method:** Lightning-managed (single `python` process, Lightning spawns workers).\n\n| Spec Key | Description | Default |\n|----------|-------------|---------|\n| `train.num_gpus` | Number of GPUs | 1 |\n| `train.gpu_ids` | GPU device indices | [0] |\n\n- Strategy: `auto` (Lightning picks best strategy automatically)\n- No explicit `num_nodes` or `distributed_strategy` config — single-node only\n- Lightweight model, single GPU typically sufficient\n\n## Hardware\n\nMinimum 1 GPU(s), recommended 1 GPU(s). 8GB+ VRAM per GPU. Pose classification is very lightweight — skeleton data is small. Single GPU is sufficient.\n\n## Error Patterns\n\n**Graph layout mismatch**: Ensure model.graph_layout matches the skeleton format in your .npy data files.\n\n**Label shape mismatch**: train_label.pkl class indices must be in range [0, num_classes).\n\n**Missing label map**: The training dataloader expects `dataset.label_map` to be a dictionary. If the dataset only supplies numeric class IDs, set a synthetic contiguous map such as `class_0: 0` through `class_5: 5` for the six-class NVIDIA sample data.\n\n**Checkpoint handoff**: After AutoML\u002Ftrain, use the checkpoint resolver to select the intended saved `.pth` checkpoint under the parent result folder, such as `model_epoch_000_step_00007.pth`, and pass that exact file as `evaluate.checkpoint`, `export.checkpoint`, `inference.checkpoint`, or `train.resume_training_checkpoint_path`. `pc_model_latest.pth` is a latest-checkpoint symlink; use it only when the user explicitly asks for latest rather than a specific\u002Fbest checkpoint. Keep the same `dataset.num_classes`, `dataset.label_map`, and `model.graph_layout` overrides for downstream actions.\n\n**Dataset conversion source**: `dataset_convert` expects the raw JSON output from the DeepStream BodyPose app. The common NVIDIA sample S3 folder is already converted to `train_data.npy`, `train_label.pkl`, `val_data.npy`, `val_label.pkl`, `test_data.npy`, and `test_label.pkl`; skip conversion and start from the converted files when those are present.\n\n**Action-specific dataset paths**: The evaluate and inference templates also contain the training `dataset.train_dataset` and `dataset.val_dataset` blocks. For evaluate, populate `evaluate.test_dataset.data_path` and `evaluate.test_dataset.label_path`. For inference, populate `inference.test_dataset.data_path` and set `inference.output_file`; do not stop after replacing the first `data_path` or `label_path` in the file.\n\n**Output files**: Export needs an explicit `export.onnx_file` path. Inference must set `inference.output_file` to a writable file path; the packaged template default is an empty string, and the current PyTorch inference code opens that value directly.\n\n## Spec Param \u002F Parent Model Inference\n\nModel-specific inference mappings belong in this MD file, not in `config.json`. Generated runners should read this section and apply the mappings with SDK helpers before `create_job()`. This mirrors the old microservices `infer_params.py` flow.\n\nInference mappings from TAO Core `pose_classification.config.json`:\n\n| Action | Spec Field | Inference Function | Meaning |\n|---|---|---|---|\n| dataset_convert | `dataset_convert.results_dir` | `output_dir` | current job results directory |\n| evaluate | `encryption_key` | `key` | encryption key |\n| evaluate | `evaluate.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| evaluate | `results_dir` | `output_dir` | current job results directory |\n| export | `encryption_key` | `key` | encryption key |\n| export | `export.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| export | `export.onnx_file` | `create_onnx_file` | output ONNX path |\n| export | `results_dir` | `output_dir` | current job results directory |\n| inference | `encryption_key` | `key` | encryption key |\n| inference | `inference.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| inference | `inference.output_file` | `create_inference_result_file_pose` | pose inference result file |\n| inference | `results_dir` | `output_dir` | current job results directory |\n| train | `encryption_key` | `key` | encryption key |\n| train | `model.pretrained_model_path` | `ptm_if_no_resume_model` | PTM when no resume checkpoint exists |\n| train | `results_dir` | `output_dir` | current job results directory |\n| train | `train.resume_training_checkpoint_path` | `resume_model` | model file inferred from the current job results folder |\n\nFor `parent_model` or `parent_model_folder`, pass the upstream train\u002Fexport\u002FAutoML child job id as `parent_job_id`. The SDK lists the parent result folder, filters checkpoint artifacts, and returns the selected model file or folder. Do not add these mappings back to `config.json` and do not patch generated runner scripts to guess checkpoint paths.\n","# Pose Classification\n\nPose classification using ST-GCN (Spatial Temporal Graph Convolutional Network). Classifies skeleton sequences into action categories from pose keypoint data.\n\nTypically trained from scratch on skeleton data.\n\nThe packaged PyTorch Pose Classification CLI supports `dataset_convert`, `train`, `evaluate`, `export`, and `inference`. `dataset_convert` is conditional: run it only when the input is raw DeepStream BodyPose JSON. If the dataset is already converted to TAO-ready `.npy` \u002F `.pkl` files, start directly with `train` on those files and mark dataset conversion as `not run: preconverted dataset provided` in validation reports. This model does not expose deploy, prune, quantize, or standalone retrain actions. Resume\u002Fretrain behavior uses `pose_classification train -e ...` with `train.resume_training_checkpoint_path` populated.\n\n## Dataclass Schemas\n\nGenerated TAO Core schemas are packaged in `schemas\u002F\u003Caction>.schema.json`, with `schemas\u002Fmanifest.json` listing available actions. Each generated schema also emits `references\u002Fspec_template_\u003Caction>.yaml` from the schema top-level `default` field. AutoML enablement is declared at the model layer in `references\u002Fskill_info.yaml` via `automl_enabled`. Runnable AutoML still requires `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` to exist and parse. Use the packaged train schema for `automl_default_parameters`, `automl_disabled_parameters`, defaults, min\u002Fmax bounds, enums, option weights, math conditions, dependencies, and popular parameters. Do not expect `~\u002Ftao-core` at runtime; maintainers regenerate schemas\u002Ftemplates before packaging the skill bank.\n\n## Train Action Policy\n\nThis model is AutoML-enabled at the model layer. Before handling any train-stage request, read `references\u002Fskill_info.yaml` and resolve the run override from either an explicit `automl_policy` value or the user's workflow request. Use `automl_policy: on` by default and only expose `on` \u002F `off` in new launch prompts. Treat phrases like \"turn off AutoML\", \"disable AutoML\", \"no HPO\", or \"plain training\" as `automl_policy: off` for this run only. When `automl_policy: on`, `automl_enabled: true`, and both `schemas\u002Ftrain.schema.json` and `references\u002Fspec_template_train.yaml` are packaged, route the train action through `tao-skill-bank:tao-run-automl` by default with this model's `skill_dir`. Preserve workflow\u002Fapplication overrides for datasets, specs, output directories, GPU\u002Fplatform settings, parent checkpoints, and `automl_policy`. Use direct model training only when `automl_policy: off` or the packaged train schema\u002Ftemplate is missing; in the missing-schema case, report that AutoML is enabled but not runnable for this model until schemas are generated.\n\nNon-train actions such as `evaluate`, `inference`, `export`, and deploy flows stay in this model skill. The per-run `automl_policy` override does not change model metadata.\n\n## Training Requirements\n\n- **Dataset type:** pose_classification\n- **Formats:** default\n- **Monitoring metric:** val_loss\n\n### Per-Action Dataset Requirements\n\n| Action | Spec Key | Source | Files | List? |\n|---|---|---|---|---|\n| dataset_convert (optional) | dataset_convert.data | id | DeepStream BodyPose JSON | No |\n| evaluate | evaluate.test_dataset.data_path | train_datasets | val_data.npy | No |\n| evaluate | evaluate.test_dataset.label_path | train_datasets | val_label.pkl | No |\n| inference | inference.test_dataset.data_path | train_datasets | test_data.npy | No |\n| train | dataset.train_dataset.data_path | train_datasets | train_data.npy | No |\n| train | dataset.train_dataset.label_path | train_datasets | train_label.pkl | No |\n| train | dataset.val_dataset.data_path | train_datasets | val_data.npy | No |\n| train | dataset.val_dataset.label_path | train_datasets | val_label.pkl | No |\n\n### Typical Spec Overrides\n\nData source overrides are **mandatory for every action being run** — the agent MUST construct data source paths from the Per-Action Dataset Requirements table above and include them in `spec_overrides`. Do not run `dataset_convert` when the supplied dataset is already converted to `.npy` \u002F `.pkl` files.\n\n```python\nS3_TRAIN = \"s3:\u002F\u002Fbucket\u002Fdata\u002Fpurpose_built_models_pose_classification_train\u002Fnvidia\"\nCHECKPOINT = \"\u002Fresults\u002F{train_job_id}\u002Fresults_dir\u002Fmodel_epoch_000_step_00007.pth\"\n```\n\n**dataset_convert (optional; raw DeepStream BodyPose JSON only):**\n```python\n{\n    \"dataset_convert.data\": \"s3:\u002F\u002Fbucket\u002Fdata\u002F\u003Cdeepstream-bodypose-output>.json\",\n}\n```\n\n**train (mandatory data sources):**\n```python\n{\n    \"train.num_epochs\": 30,\n    \"train.checkpoint_interval\": 10,\n    \"train.validation_interval\": 10,\n    \"train.num_gpus\": 1,\n    \"wandb.enable\": False,\n    \"dataset.num_classes\": 6,\n    \"dataset.label_map\": {\n        \"class_0\": 0,\n        \"class_1\": 1,\n        \"class_2\": 2,\n        \"class_3\": 3,\n        \"class_4\": 4,\n        \"class_5\": 5,\n    },\n    \"model.graph_layout\": \"nvidia\",\n    \"dataset.train_dataset.data_path\": f\"{S3_TRAIN}\u002Ftrain_data.npy\",\n    \"dataset.train_dataset.label_path\": f\"{S3_TRAIN}\u002Ftrain_label.pkl\",\n    \"dataset.val_dataset.data_path\": f\"{S3_TRAIN}\u002Fval_data.npy\",\n    \"dataset.val_dataset.label_path\": f\"{S3_TRAIN}\u002Fval_label.pkl\",\n}\n```\n\n**resume train (mandatory checkpoint):**\n```python\n{\n    \"train.num_epochs\": 31,\n    \"train.resume_training_checkpoint_path\": CHECKPOINT,\n    \"dataset.train_dataset.data_path\": f\"{S3_TRAIN}\u002Ftrain_data.npy\",\n    \"dataset.train_dataset.label_path\": f\"{S3_TRAIN}\u002Ftrain_label.pkl\",\n    \"dataset.val_dataset.data_path\": f\"{S3_TRAIN}\u002Fval_data.npy\",\n    \"dataset.val_dataset.label_path\": f\"{S3_TRAIN}\u002Fval_label.pkl\",\n}\n```\n\n**evaluate (mandatory data sources):**\n```python\n{\n    \"evaluate.test_dataset.data_path\": f\"{S3_TRAIN}\u002Fval_data.npy\",\n    \"evaluate.test_dataset.label_path\": f\"{S3_TRAIN}\u002Fval_label.pkl\",\n    \"evaluate.checkpoint\": CHECKPOINT,\n}\n```\n\n**export (mandatory checkpoint and output):**\n```python\n{\n    \"export.checkpoint\": CHECKPOINT,\n    \"export.onnx_file\": \"\u002Fresults\u002F{export_job_id}\u002Fresults_dir\u002Fpose_classification.onnx\",\n}\n```\n\n**inference (mandatory data sources):**\n```python\n{\n    \"inference.test_dataset.data_path\": f\"{S3_TRAIN}\u002Ftest_data.npy\",\n    \"inference.test_dataset.label_path\": f\"{S3_TRAIN}\u002Ftest_label.pkl\",\n    \"inference.checkpoint\": CHECKPOINT,\n    \"inference.output_file\": \"\u002Fresults\u002Fpose_classification_inference.txt\",\n}\n```\n## Dataset Convert\n\nDataset conversion is optional for Pose Classification. Run `pose_classification dataset_convert` only when the user supplies raw DeepStream BodyPose JSON. For the common S3 validation dataset, the data is already converted to `train_data.npy`, `train_label.pkl`, `val_data.npy`, `val_label.pkl`, `test_data.npy`, and `test_label.pkl`; use those files directly for train\u002Fevaluate\u002Finference\u002Fexport flows and do not synthesize fake BodyPose JSON.\n\n## Eval Dataset\n\nOptional. Validation data is provided alongside training as val_data.npy \u002F val_label.pkl. TAO training emits `val_loss` as the TensorBoard validation scalar for this model; use `val_loss` with minimize direction for AutoML selection unless a custom evaluation hook supplies a different metric.\n\n## Important Parameters\n\n- **dataset.num_classes**: Number of pose action classes. Default 6.\n- **model.graph_layout**: Skeleton graph layout. Options: nvidia, openpose. Determines joint connectivity.\n- **model.graph_strategy**: Graph partitioning strategy for GCN.\n- **train.optim.lr**: Learning rate. Default 0.1 (SGD). Higher than vision models due to graph convolution properties.\n- **model.dropout**: Dropout rate for regularization.\n\n## Multi-GPU \u002F Multi-Node\n\n**Launch method:** Lightning-managed (single `python` process, Lightning spawns workers).\n\n| Spec Key | Description | Default |\n|----------|-------------|---------|\n| `train.num_gpus` | Number of GPUs | 1 |\n| `train.gpu_ids` | GPU device indices | [0] |\n\n- Strategy: `auto` (Lightning picks best strategy automatically)\n- No explicit `num_nodes` or `distributed_strategy` config — single-node only\n- Lightweight model, single GPU typically sufficient\n\n## Hardware\n\nMinimum 1 GPU(s), recommended 1 GPU(s). 8GB+ VRAM per GPU. Pose classification is very lightweight — skeleton data is small. Single GPU is sufficient.\n\n## Error Patterns\n\n**Graph layout mismatch**: Ensure model.graph_layout matches the skeleton format in your .npy data files.\n\n**Label shape mismatch**: train_label.pkl class indices must be in range [0, num_classes).\n\n**Missing label map**: The training dataloader expects `dataset.label_map` to be a dictionary. If the dataset only supplies numeric class IDs, set a synthetic contiguous map such as `class_0: 0` through `class_5: 5` for the six-class NVIDIA sample data.\n\n**Checkpoint handoff**: After AutoML\u002Ftrain, use the checkpoint resolver to select the intended saved `.pth` checkpoint under the parent result folder, such as `model_epoch_000_step_00007.pth`, and pass that exact file as `evaluate.checkpoint`, `export.checkpoint`, `inference.checkpoint`, or `train.resume_training_checkpoint_path`. `pc_model_latest.pth` is a latest-checkpoint symlink; use it only when the user explicitly asks for latest rather than a specific\u002Fbest checkpoint. Keep the same `dataset.num_classes`, `dataset.label_map`, and `model.graph_layout` overrides for downstream actions.\n\n**Dataset conversion source**: `dataset_convert` expects the raw JSON output from the DeepStream BodyPose app. The common NVIDIA sample S3 folder is already converted to `train_data.npy`, `train_label.pkl`, `val_data.npy`, `val_label.pkl`, `test_data.npy`, and `test_label.pkl`; skip conversion and start from the converted files when those are present.\n\n**Action-specific dataset paths**: The evaluate and inference templates also contain the training `dataset.train_dataset` and `dataset.val_dataset` blocks. For evaluate, populate `evaluate.test_dataset.data_path` and `evaluate.test_dataset.label_path`. For inference, populate `inference.test_dataset.data_path` and set `inference.output_file`; do not stop after replacing the first `data_path` or `label_path` in the file.\n\n**Output files**: Export needs an explicit `export.onnx_file` path. Inference must set `inference.output_file` to a writable file path; the packaged template default is an empty string, and the current PyTorch inference code opens that value directly.\n\n## Spec Param \u002F Parent Model Inference\n\nModel-specific inference mappings belong in this MD file, not in `config.json`. Generated runners should read this section and apply the mappings with SDK helpers before `create_job()`. This mirrors the old microservices `infer_params.py` flow.\n\nInference mappings from TAO Core `pose_classification.config.json`:\n\n| Action | Spec Field | Inference Function | Meaning |\n|---|---|---|---|\n| dataset_convert | `dataset_convert.results_dir` | `output_dir` | current job results directory |\n| evaluate | `encryption_key` | `key` | encryption key |\n| evaluate | `evaluate.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| evaluate | `results_dir` | `output_dir` | current job results directory |\n| export | `encryption_key` | `key` | encryption key |\n| export | `export.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| export | `export.onnx_file` | `create_onnx_file` | output ONNX path |\n| export | `results_dir` | `output_dir` | current job results directory |\n| inference | `encryption_key` | `key` | encryption key |\n| inference | `inference.checkpoint` | `parent_model` | model file inferred from the parent job results folder |\n| inference | `inference.output_file` | `create_inference_result_file_pose` | pose inference result file |\n| inference | `results_dir` | `output_dir` | current job results directory |\n| train | `encryption_key` | `key` | encryption key |\n| train | `model.pretrained_model_path` | `ptm_if_no_resume_model` | PTM when no resume checkpoint exists |\n| train | `results_dir` | `output_dir` | current job results directory |\n| train | `train.resume_training_checkpoint_path` | `resume_model` | model file inferred from the current job results folder |\n\nFor `parent_model` or `parent_model_folder`, pass the upstream train\u002Fexport\u002FAutoML child job id as `parent_job_id`. The SDK lists the parent result folder, filters checkpoint artifacts, and returns the selected model file or folder. Do not add these mappings back to `config.json` and do not patch generated runner scripts to guess checkpoint paths.\n",{"name":5,"tags":24,"license":15,"metadata":27,"description":7,"allowed-tools":30,"compatibility":31},[25,26],"pose","classification",{"author":28,"version":29},"NVIDIA Corporation","0.1.0","Read Bash","Requires docker + nvidia-container-toolkit.","main","1c523aba5cdf867a79f078847075492d7778092a","public\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification",1,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills","https:\u002F\u002Fskills.x402.wtf","https:\u002F\u002Fgithub.com\u002Fthemusebook","https:\u002F\u002Fmusebook.trade","The official Github of musebook","2026-11-02 23:42:26",{"source":8,"skillName":5,"commands":43,"prompt":53,"zipAvailable":54},[44,47,50],{"tool":45,"command":46},"npx","npx skills add themusebook\u002Fskills --skill tao-train-pose-classification",{"tool":48,"command":49},"bunx","bunx skills add themusebook\u002Fskills --skill tao-train-pose-classification",{"tool":51,"command":52},"pnpm","pnpm dlx skills add themusebook\u002Fskills --skill tao-train-pose-classification","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002FSKILL.md",true,[56,62,66,70,74,78,82,86,90,94],{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"BENCHMARK.md","REFERENCE","参考文档",4141,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002FBENCHMARK.md",{"path":63,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":64,"sourceHtmlUrl":65},"metadata.json",629,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Fmetadata.json",{"path":67,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":68,"sourceHtmlUrl":69},"references\u002Fskill_info.yaml",4493,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Freferences\u002Fskill_info.yaml",{"path":71,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":72,"sourceHtmlUrl":73},"references\u002Fspec_template_dataset_convert.yaml",517,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Freferences\u002Fspec_template_dataset_convert.yaml",{"path":75,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":76,"sourceHtmlUrl":77},"references\u002Fspec_template_evaluate.yaml",1482,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Freferences\u002Fspec_template_evaluate.yaml",{"path":79,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":80,"sourceHtmlUrl":81},"references\u002Fspec_template_export.yaml",1374,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Freferences\u002Fspec_template_export.yaml",{"path":83,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":84,"sourceHtmlUrl":85},"references\u002Fspec_template_inference.yaml",1501,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Freferences\u002Fspec_template_inference.yaml",{"path":87,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":88,"sourceHtmlUrl":89},"references\u002Fspec_template_train.yaml",1304,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Freferences\u002Fspec_template_train.yaml",{"path":91,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":92,"sourceHtmlUrl":93},"skill-card.md",4112,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Fskill-card.md",{"path":95,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":96,"sourceHtmlUrl":97},"verification.json",2709,"https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fpublic\u002Fapi\u002Fskills\u002Fnvidia\u002Ftao-train-pose-classification\u002Fverification.json",[99,108,115,122,129,136],{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":103,"licenseFileUrl":104,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":19},"01M7AB34SPKVSDRRG009PM1CZH","youtube-clipper","YouTube 视频智能剪辑工具。下载视频和字幕，AI 分析生成精细章节（几分钟级别）， 用户选择片段后自动剪辑、翻译字幕为中英双语、烧录字幕到视频，并生成总结文案。 使用场景：当用户需要剪辑 YouTube 视频、生成短视频片段、制作双语字幕版本时。 关键词：视频剪辑、YouTube、字幕翻译、双语字幕、视频下载、clip video","MIT","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-clipper\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fyoutube-clipper\u002FSKILL.md",[],[],{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":103,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":19},"01M7AB18BH6N7KR7939HRNWVMQ","viem-integration","Integrate EVM blockchains using viem. Use when user says \"read blockchain data\", \"send transaction\", \"interact with smart contract\", \"connect to Ethereum\", \"use viem\", \"use wagmi\", \"wallet integration\", \"viem setup\", or mentions blockchain\u002FEVM development with TypeScript.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fviem-integration\u002FSKILL.md",[],[],{"skillNo":116,"name":117,"nameZh":6,"description":118,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":103,"licenseFileUrl":6,"sourceHtmlUrl":119,"categories":120,"keywords":121,"publishedAt":19},"01M7AB12GHGP48WCNXDQQ9222H","v4-security-foundations","Security-first Uniswap v4 hook development. Use when user mentions \"v4 hooks\", \"hook security\", \"PoolManager\", \"beforeSwap\", \"afterSwap\", or asks about V4 hook best practices, vulnerabilities, or audit requirements.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fv4-security-foundations\u002FSKILL.md",[],[],{"skillNo":123,"name":124,"nameZh":6,"description":125,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":103,"licenseFileUrl":6,"sourceHtmlUrl":126,"categories":127,"keywords":128,"publishedAt":19},"01M7AB0XM3PSR7V3Q2ZRR0Q0FB","v4-sdk-integration","App-layer SDK guide for building swap and liquidity experiences directly with the Uniswap v4 SDK. Use when user asks about \"v4 sdk\", \"uniswap v4\", \"v4 swap\", \"v4 liquidity\", \"PoolManager\", \"V4Planner\", \"StateView\", \"PositionManager\", \"pool state\", \"v4 position\", \"uniswap sdk\", or when building swap\u002Fliquidity UX directly with SDKs rather than via the Trading API.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fv4-sdk-integration\u002FSKILL.md",[],[],{"skillNo":130,"name":131,"nameZh":6,"description":132,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":103,"licenseFileUrl":6,"sourceHtmlUrl":133,"categories":134,"keywords":135,"publishedAt":19},"01M7AB0SGA93ASA588DJW4P9WT","v4-hook-generator","Generate Uniswap v4 hook contracts via OpenZeppelin MCP. Use when building custom swap logic, async swaps, hook-owned liquidity, custom curves, dynamic fees, MEV protection, limit orders, or oracle hooks.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fv4-hook-generator\u002FSKILL.md",[],[],{"skillNo":137,"name":138,"nameZh":6,"description":139,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":103,"licenseFileUrl":6,"sourceHtmlUrl":140,"categories":141,"keywords":142,"publishedAt":19},"01M7AAZYH75H4CS03WSW2EZ8DM","testing","For testing with Light Protocol programs and clients on localnet, devnet, and mainnet validation.","https:\u002F\u002Fgithub.com\u002Fthemusebook\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftesting\u002FSKILL.md",[],[]]