recognition
This skill provides ST-GCN skeleton action-recognition model, graph, configuration, checkpoint, and CLI guidance while routing data and pose concerns to their owning skills.
- License 许可
- Apache-2.0
- 最近更新
- 2026-10-08
- Stars
- ★ 11
- Forks
- 3
01 前言 Front Matter
- name
- recognition
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- This skill provides ST-GCN skeleton action-recognition model, graph, configuration, checkpoint, and CLI guidance while routing data and pose concerns to their owning skills.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/api-reference.md3.6 KB
- 参考文档references/cli-reference.md4.5 KB
- 参考文档references/model-zoo.md2.5 KB
- 参考文档references/troubleshooting.md6.0 KB
03 SKILL.md 正文
Recognition
Use this sub-skill when a user needs the repository's ST_GCN_18 model, graph
layout/strategy selection, recognition train/test configuration construction,
pretrained-checkpoint alias selection, mmskl argument binding, or model
input/output shape guidance.
- Read API reference for constructor contracts, graph layouts, adjacency strategies, and tensor shapes.
- Read CLI and configuration reference for recognition train/test patterns, flags, batch/GPU behavior, and config-path handling.
- Read model zoo reference for checkpoint aliases and the no-download boundary.
- Run the tiny ST-GCN smoke for a bounded,
download-free import/graph/forward check. It asserts finite output and the
requested
(N, num_class)shape; it is not training or evaluation. - Use troubleshooting for CUDA/compiler, graph/class-count, data/checkpoint, CLI, and native-NMS gates.
Scope and routing
This sub-skill covers the recognition model and its application wiring. Route JSON schema, annotation validation, loader/transforms, and axis conversion to the data-preparation skill. Route video input, MMDetection, HRNet, and pose extraction to the pose-estimation skill; successful ST-GCN construction does not establish detector or video readiness.
The smoke deliberately creates synthetic input and never downloads data or a checkpoint. Do not claim a complete training/evaluation run or downloaded checkpoint verification from this sub-skill.
04 同仓库更多 Skills
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
python-api
Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP/mAP semantics, and source-style import diagnostics.
file-evaluation
Folder-based PASCAL VOC AP/mAP evaluation from ground-truth and detection text files.
object-detection-metrics
Guides legacy Object-Detection-Metrics PASCAL VOC AP/mAP evaluation from detection text folders and direct Python API classes.
nx-workspace-patterns
Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.