skills.rest
rilical
rilicalCOMMUNITY
Explore all skills by this author →
GitHub

baseline-detector

Integrate the CLIP baseline detector with a frozen encoder and optional residual fusion.

Updated Feb 18, 2026
Search Tags:#config#inference#clip#fusion#baseline-detector#ai-real
One-click install
npx skills add https://github.com/rilical/OpenWorld-AI-Image-Detection --skill baseline-detector
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: baseline-detector
Source: https://github.com/rilical/OpenWorld-AI-Image-Detection/tree/main/.agents/skills/baseline-detector
Command: npx skills add https://github.com/rilical/OpenWorld-AI-Image-Detection --skill baseline-detector

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps maintain and extend the CLIP baseline detector by providing guidance for stable integration, optional fusion paths, and clear detector-side inference contracts.

Core Features & Use Cases

  • Baseline detector path with encoder frozen by default
  • Optional residual fusion modules and direction-fusion knobs
  • Clear detector-side inference contracts for AI/Real outputs and easy testing

Quick Start

Tune the CLIP baseline detector by enabling the frozen encoder and optional residual fusion in your config, then run the detector.

Frequently Asked Questions about baseline-detector

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I keep the CLIP encoder frozen while running inference for AI and Real detection?▼

To keep the CLIP encoder frozen during inference, apply the baseline detector configuration that locks the encoder by default and exposes two-class logits for Real and AI outputs.

Can I add residual fusion to the CLIP baseline detector without changing the default inference path?▼

Yes, you can wire optional residual fusion modules and direction-fusion knobs into the CLIP baseline detector while preserving a clean baseline path for standard inference.

How does the CLIP baseline detector output AI and Real classification scores?▼

The CLIP baseline detector outputs clear two-class logits representing Real and AI predictions, maintaining a strict detector-side inference contract for easy testing and evaluation.

Does the CLIP baseline detector handle dataset loading or evaluation plotting?▼

No, the CLIP baseline detector remains completely isolated from dataset loading and plotting code, focusing exclusively on inference workflows and detector configuration knobs.

What configuration knobs are available for tuning the CLIP baseline detector?▼

The detector supports configuration knobs for enabling the frozen encoder, toggling optional residual fusion modules, and adjusting direction-fusion settings during inference and evaluation workflows.

Related Skills

Explore More Software Engineering Skills→
ovachieverovachiever
52

clip

Encode images and texts to compute cosine similarity for zero-shot classification and cross-modal retrieval.

Community
Advanced
flashinfer-aiflashinfer-ai
264

add-reference-tests

Generate and validate pytest tests for flashinfer_trace reference implementations against FlashInfer or SGLang ground truth.

Official
Advanced
leoplastureleoplasture

paper-plot

Generate publication-quality figures from numeric data using Python scripts.

Community
Advanced
SnakeOSnakeO
9

debug:angular

Diagnose Angular runtime errors using DevTools, ng.probe, and console utilities.

Community
Advanced