---

name: model-and-backend-selection description: "Select and troubleshoot DeepFace models, detector backends, optional dependencies, model weights, TensorFlow/Keras compatibility, normalization, and encryption." disable-model-invocation: true metadata: disco-role: operating license: MIT

Model And Backend Selection

Use this sub-skill when the user task needs Supported recognition, demography, spoofing, and detector names; DeepFace.build_model; base versus optional dependencies; TensorFlow/Keras/tf-keras; weight cache; CPU/GPU; normalization and encrypted embeddings.

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

Detailed task workflows route to sibling sub-skills; database clients and services go to ../datastore-search/SKILL.md; REST startup goes to ../api-service/SKILL.md.

First Checks

1. Confirm from deepface import DeepFace works; otherwise use ../../references/troubleshooting.md.
2. Identify whether the user wants static guidance, code generation, environment diagnosis, or runtime execution.
3. Keep network/model-weight/database/camera side effects explicit and ask before running them in constrained environments.
4. Use the nearest bundled helper for safe validation before recommending a heavier run.

Reference Map

• references/api-reference.md documents signatures, parameters, outputs, and API-specific behavior for this sub-skill.
• references/workflows.md or the focused workflow reference contains practical recipes and decision patterns.
• references/troubleshooting.md maps common errors to concrete recovery steps.
• scripts/inspect_deepface_models.py is 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.