recognition-workflows
Use DeepFace verification, embeddings, local folder face search, distance metrics, thresholds, and signed local datastores.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- recognition-workflows
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Use DeepFace verification, embeddings, local folder face search, distance metrics, thresholds, and signed local datastores.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/api-reference.md2.4 KB
- 参考文档references/troubleshooting.md1.3 KB
- 参考文档references/workflows.md2.0 KB
- 脚本scripts/deepface_embedding_distance.py2.8 KB
03 SKILL.md 正文
Recognition Workflows
Use this sub-skill when the user task needs DeepFace.verify, DeepFace.represent, DeepFace.find, distance metrics, thresholds/confidence, precomputed embeddings, batch inputs, local pickle datastores, and signed LightDSA datastores.
Route Here
- Answer package-user questions for the owned DeepFace functions and parameters.
- Select safe minimal code patterns before running heavy model builds, weight downloads, database connections, Docker, or webcam/video operations.
- Diagnose workflow-specific errors using the linked troubleshooting reference.
- Cross-link to sibling sub-skills when a workflow spans detection, recognition, persistence, serving, and optional model/dependency setup.
Reroute
Face detector choice and demography details go to ../detection-and-demography/SKILL.md; database-backed register/search/build_index go to ../datastore-search/SKILL.md; optional model dependency and weight failures go to ../model-and-backend-selection/SKILL.md; REST calls go to ../api-service/SKILL.md.
First Checks
- Confirm
from deepface import DeepFaceworks; otherwise use../../references/troubleshooting.md. - Identify whether the user wants static guidance, code generation, environment diagnosis, or runtime execution.
- Keep network/model-weight/database/camera side effects explicit and ask before running them in constrained environments.
- Use the nearest bundled helper for safe validation before recommending a heavier run.
Reference Map
references/api-reference.mddocuments signatures, parameters, outputs, and API-specific behavior for this sub-skill.references/workflows.mdor the focused workflow reference contains practical recipes and decision patterns.references/troubleshooting.mdmaps common errors to concrete recovery steps.scripts/deepface_embedding_distance.pyis the safe bundled helper for this sub-skill.
Safety
Do not tell future agents to open or run original repository tests, examples, notebooks, scripts, or local checkout files. Use this self-contained sub-skill, its references, and its bundled helper scripts instead.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).