Skip to main content
Ctrl+K
Logo image

Getting Started

  • Installation
  • Package Structure
  • Quick Start
  • Notebooks
    • API
      • Dataset Provider
      • Trainer (Base)
    • Discrete Search
      • Search Spaces
      • Evaluators
      • Algorithms
      • Configuration-based Search
    • Computer Vision
      • Dataset Provider
      • PyTorch-Lightining Trainer
    • Natural Language Processing
      • Fast HF Dataset Provider
      • HF Dataset Provider
      • HF Trainer
      • NVIDIA Dataset Provider
      • NVIDIA Trainer
      • ONNX Export
      • PyTorch Quantization
      • Transformer++ Search Space

Advanced Guide

  • Cloud-Based Search
    • Azure
      • Notebooks
        • Quickstart
        • Text Generation
        • Multi node search

Contributing

  • First Time Contributor
  • Documentation
  • Unit Tests

Support

  • Frequently Asked Questions
  • Contact
  • Copyright

Reference

  • API
    • API
    • Common Packages
    • Datasets
      • Computer Vision
        • Transforms
      • Natural Language Processing
        • Tokenization Utilities
    • Discrete Search
      • Search Algorithms
      • API
      • Evaluators
        • Benchmark
        • Natural Language Processing
        • PyTorch Profiler (Utilities)
      • Predictors
      • Search Spaces
        • Benchmark
        • Configuration-Based
        • Computer Vision
        • Natural Language Processing
      • Utilities
    • ONNX
      • Configuration Utilities
      • Optimization Utilities
    • Quantization
      • Natural Language Processing
    • Supergraph
      • Algorithms
        • DARTS
        • DiDARTS
        • DivNAS
        • Gumbel-Softmax
        • Manual
        • NasBench-101
        • Petridish
        • Random
        • XNAS
      • Datasets
        • Providers
      • Models
        • ShakeShake
      • Neural Architecture Search
      • Utilities
    • Trainers
      • Computer Vision
      • Natural Language Processing
  • Roadmap
  • Changelog
  • Repository
  • Open issue

Discrete Search

Discrete Search#

  • Search Spaces
    • The ArchaiModel class
    • Building a Search Space
    • Making the search space compatible with NAS algorithms
    • Built-in Search Spaces
  • Evaluators
    • Evaluating models
  • Algorithms
    • Dataset Provider
    • Wrapping custom evaluation code
    • Defining Search Objectives
    • Using a search algorithm
  • Configuration-based Search
    • Creating an ArchParamTree
    • More features of ArchParamTrees
    • Example: Building an Image Classification Search Space
    • Tracking used architecture parameters for model de-duplication

previous

Implementing a Custom Trainer

next

Discrete Search Spaces

By Microsoft

© Copyright 2023.

Last updated on Apr 27, 2023.