Skip to content

YOLO v3 TensorFlow Lite iOS GPU acceleration. Training from scratch and making a GPU accelerated mobile application.

Notifications You must be signed in to change notification settings

JeiKeiLim/tflite-yolov3-gpu-ready

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 

Repository files navigation

YOLO v3 TensorFlow Lite iOS GPU acceleration

  • This projects trains YOLO v3 model with BDD100k Dataset
  • And you can convert the trained model into TFLite model that GPU acceleration ready for iOS!

Training

Please refer to python directory of this repository.

iOS Examples

You can find out iOS directory of this repository.

1. YOLO object detection example project

Please refer to iOS/yolojk_iOS directory of this repository.

iPhone X iPadPro 11"

2. gonogono - Observing traffic signs and front car instead of the driver

Please refer to iOS/gonogono directory of this repository.

Performance

Model Name # Params # Label Device FPS
MobileNetV1 ~9M 4 iPhone X 15~20
MobileNetV1 ~2.3M 4 iPhone X 30~35
MobileNetV1 ~2.3M 4 iPad Pro 11" 60~70

Future update

  • Make model smaller -> under 5MB
  • Int8 quantization support
  • Android example application
  • CoreML Support

About

YOLO v3 TensorFlow Lite iOS GPU acceleration. Training from scratch and making a GPU accelerated mobile application.

Topics

Resources

Stars

Watchers

Forks

Languages