English

How to Turn on AMP

When AMP (Automatic Mixed Precision) is enabled, Pytorch can automatically compute some operators (such as convolution and full concatenation) by using float16 during model execution to increase computing speed and reduce memory usage. See the Official Pytorch Documentation for more details.

AMP is prepared and ready to use in HAT. Users just need to set the enable_amp to True when defining the batch_processor field in the config file.

Note

In order to get accurate metrics during model validation, it is not necessary to turn on the AMP. To turn it off, set the enable_amp parameter to False when defining the val_batch_processor field.

# configs/example.py

# Use of BasicBatchProcessor
batch_processor = dict(
    type='BasicBatchProcessor',
    need_grad_update=...,
    batch_transforms=...,
    enable_amp=True,
)

# Use of MultiBatchProcessor
batch_processor = dict(
    type="MultiBatchProcessor",
    need_grad_update=...,
    batch_transforms=...,
    loss_collector=...,
    enable_amp=True,
)

PTQ Conversion Tools

The hb_compile Tool

PTQ Conversion Steps

PTQ Conversion Samples

FAQ and Common Failure Resolutions

Appendix

Tutorial

Advanced Tutorial

API Reference

QAT

Model Export

Horizon Operator

Vision Process

Vision Algorithm

Sample

VP API Introduction

Data Structure

Function Interface

Model Inference Development

Model Inference API Introduction

Data Type and Data Structure

API Interface

Model Inference Tool Introduction

High Performance Library

High Performance Algorithm

HPL API Introduction

Data Structure

Functional Interface

UCP Common API Introduction

Data Structure

Functional Interface

UCP Performance Analysis Tools

Horizon Torch Samples

FRAMEWORK

TUTORIALS

EXAMPLES

API REFERENCE

Toolchain Operator Support Constraint List

How to Turn on AMP

The hb_compile Tool

QAT

Model Export

Horizon Operator

Vision Algorithm

Sample

VP API Introduction

Data Structure

Function Interface

Model Inference API Introduction

Data Type and Data Structure

API Interface

Model Inference Tool Introduction

High Performance Algorithm

HPL API Introduction

Data Structure

Functional Interface

Data Structure

Functional Interface

FRAMEWORK

TUTORIALS

EXAMPLES

API REFERENCE

#How to Turn on AMP

How to Turn on AMP