FeiYull/TensorRT-Alpha

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1460 stars · 200 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A C++/CUDA library providing TensorRT-accelerated deployment for 30+ popular computer vision models including YOLOv3-v8, YOLOv8-Pose/Seg/Cls, YOLOX, YOLO NAS, EfficientDet, RT-DETR, U2Net, and libfacedetection. It implements multi-batch CUDA preprocessing, inference, decoding, and NMS on GPU, with model conversion via torch->onnx->tensorrt.

Use cases

• deploy yolov8 with tensorrt in c++
• accelerate yolo inference on nvidia gpu
• run object detection models fast with cuda
• convert pytorch models to tensorrt engines
• deploy yolov8-pose and yolov8-seg on gpu
• batch preprocess and run nms on cuda
• deploy detection models on ubuntu or windows with tensorrt

When to choose

• you need maximum GPU inference speed for YOLO-family or other CV models in C++
• you deploy on NVIDIA hardware with TensorRT and want CUDA preprocessing/decoding/NMS included
• you need pose estimation, segmentation, or detection variants of YOLOv8 in one deployment library

When to avoid

• you want a simple Python-only inference pipeline
• you have no NVIDIA GPU or cannot install CUDA/TensorRT
• you need models outside the supported CV detection/segmentation set
• you need a permissively licensed dependency (it is GPL-2.0)

Facets

library · maturity active

machine-learning deep-learning image-processing computer-vision gpu-computing llm-inference computer-vision deep-learning machine-learning gpu-computing image-processing windows cpp tensorrt cuda yolo object-detection model-deployment inference-optimization nvidia linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="FeiYull/TensorRT-Alpha")