ARCADE

Quickstart

  • Quickstart
    • Overview of ARCADE
      • What problem ARCADE solves
      • What ARCADE is not
      • Repository surfaces you will use
      • Design principles
      • Where to go next
    • Installation
      • Requirements
      • Create an environment and install
      • Optional extras
      • Command line entry point
      • Building this documentation
      • Next
    • Pipeline stages
      • Data preparation
      • Classification
      • Optional classifier optimization
      • Hardware estimation
      • Optional quantum error correction curves
      • Visualization and report generation
      • Mapping stages to packages
      • Next
    • Verification
      • Import and console script
      • Describe a baseline method
      • Smoke-run a tiny configuration
      • Link paper data when needed
      • Optional dependency checks
      • What success looks like
    • Suggested reading order

Guides

  • How-to guides
    • Compare classifiers
      • Write a minimal configuration
      • Validate, then run
      • Keep the recipe frozen
      • What fair means in practice
      • Choose smoke, zoo, or paper fidelity
      • After the run
    • Build your own classifier
      • Contract you must satisfy
      • Pattern A: nearest-mean on filter features
      • Pattern B: transformer-style custom class
      • What you get for free
      • Pitfalls
    • Enable optimization
      • See it in action
      • Flip the switch in YAML
      • What runs
      • Caveats
    • Enable QEC and FTQC curves
      • YAML configuration
      • Caveats
    • Choose a guide by goal
      • Build your own classifier
  • Explanation guides
    • What is readout?
      • What ARCADE is
      • Pipeline intuition
    • Metrics
      • Joint accuracy
      • Infidelity
      • F5Q
      • PQ min and per-qubit fidelity
      • Hardware columns: LUT, DSP, and latency
      • Duration and sweep curves
    • Shared recipe
      • What is frozen
      • What is allowed to vary
      • Pitfalls checklist
    • Hardware limits
      • What ARCADE estimates
      • How to use the columns
    • FTQC bridge
      • The mapping in one line
      • Wall-clock cycle
      • What this stage is and is not
    • Why explanation pages exist

API Reference

  • API Reference
    • CLI reference
      • arcade run
      • arcade list
      • arcade describe
      • arcade validate
      • arcade template
      • arcade explore
      • Related Python API
    • Config schema
      • Top-level sections
      • data highlights
      • classifiers
      • sweep, hardware, and visualization
      • Validate and run
    • Classifiers reference
      • Support and maturity levels
      • Reference eight-method benchmark
      • Method detail: how each reference method moves through the pipeline
      • Paper and extended recipes
      • Classical extras on filter features
      • Experimental and special names
      • Feature kinds at a glance
      • Pitfalls checklist
      • Extend
    • Classifier optimization
      • Public surface
      • Config models
      • Example
    • API modules
      • arcade
      • arcade.api
      • arcade.classifiers
      • arcade.classifieroptimization
      • arcade.config
      • arcade.data
      • arcade.features
      • arcade.hardware
      • arcade.qec
      • arcade.stages
      • arcade.runner

Tutorials

  • Tutorials
    • First zoo run
      • Install the extras used by the public zoo
      • Run the public zoo
      • Point ARCADE at your own traces
      • Next
    • Read the report
      • Where to look
      • Anatomy of report detail tiers
      • How to read the comparison table
      • Pitfalls checks
      • Next
    • Notebook: data stage in depth
      • Pipeline position
      • Expected log pattern for a cold start
      • Debugging checklist for Stage 1
      • Next
    • Notebook: features and classify stage in depth
      • Pipeline position
      • Inspect FEATURE_KIND before fitting
      • Expected log pattern
      • Debugging checklist for Stage 2
      • Next
    • Notebook: hardware estimation and report stage
      • Pipeline position
      • Expected log pattern
      • What to open after the run
      • Debugging checklist
      • Next
    • Notebook: debugging the ARCADE pipeline
      • Mental model
      • Log levels that matter
      • Symptom → stage → fix
      • Minimal reproduction template
      • Related tutorials
    • Getting started
    • Detailed notebooks
ARCADE
  • Search

© Copyright 2026, ARCADE Contributors.

Built with Sphinx using a theme provided by Read the Docs.