Skip to content

hardikdava/EyeQ

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

EyeQ

EyeQ is a minimal computer vision inference pakackge. Currently it supports following detectors in onnx runtime. It works with minimal dependencies. It is designed in such a manner to run on edge devices also.


Object Detection:

Detector onnx
yolov5
yolov6
yolov7
yolov8
yolov5u
yoloX
Damo-yolo

Instance Segmentation:

Detector Name onnx
yolov5
yolov7 #TODO
yolov8

Multi Object Tracker:

Tracker Name Integration
SORT
ByteTrack
OcSort
Norfair -

Installation:

Installation can be done via pip using following argument

 pip3 install git+https://github.com/hardikdava/EyeQ.git

TODO:

  • Docker support
  • RestAPI server ✅
  • Multi object trackers ✅
  • Instance segmentation ✅
  • Yolo Dataset loading ✅
  • COCO dataset loading
  • Object detection evaluation ✅
  • Multi object tracker evaluation ✅
  • Automatic annotation support using clip, grounding dino and sam
  • Introduce SAHI technique

Available APIs:

  • Object Detection Inference using ONNX runtime
  • Object Detction Evaluation API
  • Model serving using RESTAPI using FastAPI based server
  • Multi object Tracking for bounding boxes
  • Multi object Tracking
  • Instance segmentation support
  • Data loading for yolo

Note: models are trained using notebooks prepared by roboflow but models are not included with codebase.

References:

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages