Olive

OVERVIEW

  • Olive
  • Quick Tour
  • Olive Options

GET STARTED

  • Installation
  • Quickstart Examples

EXAMPLES

  • Examples

OLIVE COMMANDS

  • Run
  • Finetune
  • Auto-Optimization
  • Quantization
  • Capture Onnx Graph
  • Generate Adapters
  • Convert Adapters
  • Tune OnnxRuntime Session Params
  • Qualcomm SDK
  • AzureML
  • Shared Cache
  • Providing Input Models

FEATURES

  • Azure ML integration
  • Huggingface Integration
  • LoRA Adapters
  • Packaging Olive artifacts
  • Shared Cache
  • Remote Workflow
  • Model Conversions
    • PyTorch
    • ONNX
  • Model Quantizations
  • Model Transformations and Optimizations

EXTENDING OLIVE

  • Design
  • How to add new optimization Pass
  • Custom Scripts

TUTORIALS

  • How To Configure System
  • How To Configure Metric
  • How To Configure Auto Optimizer
  • How To Configure Pass
  • How To Configure Data
  • How To Set Model Path
  • Advanced User Tour
  • Self-hosted Kubernetes cluster
  • Azure ML scripts

API REFERENCE

  • OliveModels
  • ResourcePath
  • OliveSystems
  • OliveEvaluator
  • Metric
  • SearchAlgorithms
  • Engine
  • Passes
Olive
  • Model Conversions
  • View page source

Model Conversions

Olive supports converting a model from one form to another in various ways.

  • PyTorch
    • TorchTRTConversion
  • ONNX
    • Model Conversion
    • Float16 Conversion
    • Inputs/Outputs Float16 to Float32 Conversion
    • Mixed Precision Conversion
    • Convert dynamic shape to fixed shape
Previous Next

© Copyright 2023, olivedevteam@microsoft.com.

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