executing-plans
executing-plans
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available
Explore skills
Productivity · SKILL.md
saliency-and-cam
saliency-and-camRoutes users who need saliency, guided or rectified saliency, Grad-CAM, custom-loss saliency, or regression attention workflows for keras-vis.
Independent GitHub snapshot · Automated classification · Original documentation governs
Skill documentation
GitHubname: saliency-and-cam description: "Routes users who need saliency, guided or rectified saliency, Grad-CAM, custom-loss saliency, or regression attention workflows for keras-vis." disable-model-invocation: true metadata: disco-role: operating license: MIT
Saliency and CAM
Use this sub-skill when the user wants input-space attention maps from a trained keras-vis model.
Covers
visualize_saliencyvisualize_saliency_with_lossesvisualize_camvisualize_cam_with_losses- regression attention with
grad_modifier - guided or rectified saliency with
backprop_modifier - penultimate-layer selection for Grad-CAM
Routes out
- Activation image synthesis belongs in
../activation-maximization/SKILL.md. - New loss/optimizer internals belong in
../optimization-building-blocks/SKILL.md. - Heatmap overlay, color conversion, and image composition belong in
../image-utilities/SKILL.md.
What to read first
references/api-reference.mdfor exact signatures, defaults, shapes, and modifiers.references/workflows.mdfor classification, regression, and Grad-CAM recipes.references/troubleshooting.mdfor penultimate-layer, modifier, and shape recovery.
Smoke check
Use scripts/saliency_smoke.py to build a tiny deterministic model and print saliency and optional CAM heatmap shapes and value ranges.
Practical rule
Prefer saliency when the model has no close convolutional or pooling layer before the target. Use CAM when a nearby spatial layer exists and you want a class-localized spatial explanation.