wildlife-classification

What problem does it solve?

Wildlife classification tasks often require coordinating multiple model families, data pipelines, and evaluation artifacts. This Skill provides a self-contained, modular 7-agent workflow to generate training code, explanations, evaluations, and course materials for wildlife datasets.

Core Features & Use Cases

• 7-Agent pipeline orchestration: config_agent, dataset_prep_agent, model_selection_agent, fine_tuning_strategy_agent, training_code_agent, evaluation_agent, exercise_designer_agent.
• Model families supported: timm (DINOv2 backbones), DeepFaune backbone transfer, and SpeciesNet inference baselines.
• Outputs and templates: code generation, concept explanations, model evaluation, exercise design, full course modules, and model comparison reports.
• Trigger-driven discovery: responds to keywords like timm, wildlife classification, DeepFaune, SpeciesNet, fine-tune, catastrophic forgetting, and transfer learning.
• Educational tooling: supports course-module generation and guided exercises for wildlife AI literacy.

Quick Start

Configure and run a full wildlife-classification workflow to compare fine-tuning strategies across timm, DeepFaune, and SpeciesNet.