Start

Overall Structure
Models

Quick Start

Installation
Calling Source Code to Start

Usage Tutorials

Development Migration
Pretraining
Supervised Fine-Tuning (SFT)
Parameter-Efficient Fine-Tuning (PEFT)
Evaluation
Inference
MindIE Service Deployment
Quantization

Function Description

Weight Format Conversion
Distributed Weight Slicing and Merging
Distributed Parallelism
Dataset
Weight Saving and Resumable Training

Precision Optimization

Large Model Accuracy Optimization Guide

Performance Optimization

Large Model Performance Optimization Guide

API

mindformers
mindformers.core
mindformers.dataset
mindformers.generation
mindformers.models
mindformers.modules
- Operator-level parallel configuration
  - mindformers.modules.OpParallelConfig
mindformers.pet
mindformers.pipeline
mindformers.tools
mindformers.wrapper

Appendix

Environment Variable Descriptions
Configuration File Descriptions

FAQ

Model-Releated
Function_Releated
MindFormers Contribution Guidelines
Modelers Contribution Guidelines

RELEASE NOTES

Release Notes

MindSpore

»
mindformers.modules
View page source

mindformers.modules

Operator-level parallel configuration

mindformers.modules.OpParallelConfig

Configuration for operator parallelism, used to set the method of operator-level parallelism.

Previous Next

© Copyright MindSpore.

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