← Back to Projects Detection using YOLO
Real-time object detection with YOLOv8 and accelerated inference exports.
Categories
CVML
Tech Used
YOLOv8PyTorchOpenCVfine-tuningUltralyticsTensorRTONNXCUDANumPyFastAPIFlaskDocker
Problem
Accurate detection with low latency needed for real-time scenarios.
Approach
- Trained YOLOv8 with augmentation and evaluation loop
- Optimized inference using ONNX/TensorRT/CUDA
- Served via API and containerized for deployment
Results
- Low-latency detection
- Deployable artifacts (ONNX/TensorRT)
Demo Videos