Skip to content
E Elston Tan AI Scientist
Home → Projects → Talks → Achievements → Certifications →

// Projects

TensorRT Deployment For Defect Segmentation Model

Wed Jan 01 2025 Deep LearningDeploymentC++
Link to Code
tensorrt

Project Brief

Created scripts for a C++ SDK to deploy a defect segmentation model for real-time usage and compare performance/speed trade-off between different precisions

Accomplishments

  • Implemented a script to double confirm performance (mIOU) of TensorRT engine with reference to original model (from pth) performance
  • Added support for FP32/FP16/INT8 precision conversion from onnx to TensorRT engine
  • Compared performance/speed trade-off between different precisions and modes of deployment

Gallery

inference Visual confirmation with reference to original material with defect and comparison between original and TensorRT models

tensorrt-table Table for performance/speed trade-off between different precisions and modes of deployment (not all experiments included due to confidentiality)

Back