Semantic Segmentation with OpenVINO
- 0 Collaborators
Implemented inference of segmentation network with oneAPI. Compressed the model with NNCF for openVINO. ...learn more
Project status: Published/In Market
oneAPI, Internet of Things, Artificial Intelligence
Intel Technologies
oneAPI,
Intel Opt ML/DL Framework,
OpenVINO
Overview / Usage
This project lays down a template structure to do any end-to-end deployment for any sort of segmentation task. One of the best cases for this, is to deploy this onto drones, and segment forest areas, or oceans.
Methodology / Approach
Torch - Training of the model. And using torch for initial baseline to start with.
TorchScript - Transitioning to TorchScript enables dynamic Just-In-Time (JIT) compilation, enhancing computational efficiency.
Intel PyTorch Extension (IPEX) - BFFloat - IPEX with BFFloat optimization leverages oneDNN. IPEX applies graph fusion, which is accelerated by oneDNN.
Testing Quantization - Quantization with Callibration data was used then to compress the model to INT8 without compermising on accuracy.
OpenVINO Integration - The optimized model seamlessly converts to OpenVINO's Intermediate Representation (IR) format, delivering fast inference speed.
Technologies Used
- Pytorch
- OpenCV
- Pytorch Segmentation Models
- NNCF (Neural Network Compression Framework for enhanced OpenVINO™ inference)
- OpenVINO
- Intel Extention for Pytorch
Hardware - Local Laptop
Intel i7 11800H
Nvidia RTX 3060