Transfer Learning Toolkit

Introduction

• Overview
  • NVIDIA Transfer Learning Toolkit
  • Pre-trained Models
  • TLT Workflow Overview
• Requirements and Installation
  • Hardware Requirements
  • Software Requirements
  • Installation Prerequisites
  • Installation
• Supported Model Architectures
  • Model Requirements
• Purpose-Built Models
  • Training
  • Deployment
• QAT and AMP for Training

TLT Workflow

• Augmenting a Dataset
  • Configuring the Augmentor
  • Running the Augmentor Tool
• Preparing the Input Data Structure
  • Data Input for Classification
  • Data Input for Object Detection
  • Conversion to TFRecords
  • Data Input for Instance Segmentation
• Creating an Experiment Spec File
  • Specification File for Classification
  • Specification File for DetectNet_v2
  • Specification File for FasterRCNN
  • Specification File for SSD
  • Specification File for DSSD
  • Specification File for RetinaNet
  • Specification File for YOLOv3
  • Specification File for MaskRCNN
• Training the Model
  • Quantization Aware Training
  • Automatic Mixed Precision
  • Training a classification model
  • Training a DetectNet_v2 model
  • Training a FasterRCNN Model
  • Training an SSD Model
  • Training a DSSD Model
  • Training a YOLOv3 Model
  • Training a RetinaNet model
  • Training a MaskRCNN Model
• Evaluating the Model
  • Evaluating a Classification Model
  • Evaluating a DetectNet_v2 Model
  • Evaluating a FasterRCNN Model
  • Evaluating an SSD Model
  • Evaluating a DSSD model
  • Evaluating a YOLOv3 Model
  • Evaluating a RetinaNet Model
  • Evaluating a MaskRCNN Model
• Using Inference on a Model
  • Running Inference on a Classification Model
  • Running Inference on a DetectNet_v2 Model
  • Running Inference on a FasterRCNN Model
  • Running Inference on an SSD Model
  • Running Inference on a DSSD Model
  • Running Inference on a YOLOv3 Model
  • Running Inference on a RetinaNet Model
  • Running Inference on a MaskRCNN Model
• Pruning the Model
  • Required Arguments
  • Optional Arguments
  • Using the Prune Command
  • Re-training the Pruned Model
• Exporting the Model
  • INT8 Mode Overview
  • FP16/FP32 Model
  • Generating an INT8 tensorfile Using the tlt-int8-tensorfile Command
  • Exporting the Model Using tlt-export
  • INT8 Export Mode Required Arguments
  • INT8 Export Optional Arguments
  • Exporting a Model
• Deploying to Deepstream
  • TensorRT Open Source Software (OSS)
  • Generating an Engine Using tlt-converter
  • Integrating the model to DeepStream

More Information

• Release Notes
  • Transfer Learning Toolkit for Intelligent Video Analytics V2.0
• Frequently Asked Questions
  • Model support
  • Pruning
  • Model Export and Deployment
  • Training