Train with MindSpore

Quick Start

• Implementing an Image Classification Application
• Implementing Simple Linear Function Fitting
• Hands-on Installation and Experience

Basic Use

• Loading Dataset
• Defining the Network
• Saving Models
• Loading a Model for Inference and Transfer Learning
• Publishing Models using MindSpore Hub

Process Data

• Converting Dataset to MindRecord
• Optimizing Data Processing

Build Networks

• Custom Operator
• Migrating Training Scripts from Third Party Frameworks

Model Optimization

• Debugging in PyNative Mode
• Custom Debugging Information
• Training Process Visualization
• Auto Augmentation
• Evaluating the Model during Training

Performance Optimization

• Distributed Training
• Enabling Mixed Precision
• Enabling Graph Kernel Fusion
• Applying Gradient Accumulation Algorithm

Model Compression

• Applying Quantization Aware Training

Model Security and Privacy

• Improving Model Security with NAD Algorithm
• Protecting User Privacy with Differential Privacy Mechanism
• Testing Model Security Using Fuzz Testing

Application

• Computer Vision
• Natural Language Processing