Olive
0.2.0
OVERVIEW
Olive
Design
Quick Tour
Olive Options
GET STARTED
Installation
Quickstart Examples
TUTORIALS
Configuring OliveSystem
Configuring Metric
Configuring Pass
Configuring HW-dependent optimizations
ONNX related – General
PyTorch related – General
OpenVINO related – Intel HW
SNPE related – Qualcomm HW
Advanced User Tour
How to add new Pass
How to write
user_script
Packaging Olive artifacts
EXAMPLES
Inception model optimization on Qualcomm NPU
Cifar10 optimization with OpenVINO for Intel HW
BERT optimization with QAT Customized Training Loop on CPU
ResNet optimization with QAT Default Training Loop on CPU
ResNet optimization with QAT PyTorch Lightning Module on CPU
SqueezeNet latency optimization with DirectML
Stable Diffusion optimization with DirectML
BERT optimization with Intel® Neural Compressor Post Training quantization on CPU
Whisper optimization using ORT toolchain
API REFERENCE
OliveModels
OliveSystems
OliveEvaluator
Metric
SearchAlgorithms
Engine
Passes
Olive
Configuring HW-dependent optimizations
View page source
Configuring HW-dependent optimizations
¶
ONNX related – General
Model Conversion
Model Optimizer
ORT Transformers Optimization
Append Pre/Post Processing Ops
Insert Beam Serch Op
Post Training Quantization (PTQ)
ORT Performance Tuning
Float16 Conversion
Mixed Precision Conversion
PyTorch related – General
Quantization Aware Training
OpenVINO related – Intel HW
Prerequisites
Model Conversion
Post Training Quantization (PTQ)
SNPE related – Qualcomm HW
Prerequisites
Model Conversion
Post Training Quantization (PTQ)