https://github.com/alexeyab/yolo2_light

Light version of convolutional neural network Yolo v3 & v2 for objects detection with a minimum of dependencies (INT8-inference, BIT1-XNOR-inference)
https://github.com/alexeyab/yolo2_light

bit1-xnor-inference computer-vision machine-learning neural-network object-detection

Last synced: about 22 hours ago
JSON representation

Light version of convolutional neural network Yolo v3 & v2 for objects detection with a minimum of dependencies (INT8-inference, BIT1-XNOR-inference)

• Host: GitHub
• URL: https://github.com/alexeyab/yolo2_light
• Owner: AlexeyAB
• License: mit
• Created: 2017-07-18T17:37:43.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2019-08-29T10:00:48.000Z (over 5 years ago)
• Last Synced: 2024-12-31T23:08:57.493Z (8 days ago)
• Topics: bit1-xnor-inference, computer-vision, machine-learning, neural-network, object-detection
• Language: C
• Homepage:
• Size: 2.02 MB
• Stars: 304
• Watchers: 25
• Forks: 118
• Open Issues: 64
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# yolo2_light

Light version of convolutional neural network Yolo v3 & v2 for objects detection with a minimum of dependencies (INT8-inference, BIT1-XNOR-inference)

This repository supports:

* both Windows and Linux

* both OpenCV <= 3.3.0 and OpenCV 2.4.13

* both cuDNN >= 7.1.1

* CUDA >= 8.0

How to compile:

* To compile for CPU just do `make` on Linux or build `yolo_cpu.sln` on Windows

* To compile for GPU set flag `GPU=1` in the `Makefile` on Linux or build `yolo_gpu.sln` on Windows

    

    Required both [CUDA >= 8.0](https://developer.nvidia.com/cuda-toolkit-archive) and [cuDNN >= 7.1.1](https://developer.nvidia.com/rdp/cudnn-archive)

How to start:

* Download [`yolov3.weights`](https://pjreddie.com/media/files/yolov3.weights) to the `bin` directory and run `./yolo.sh` on Linux (or `yolo_cpu.cmd` / `yolo_gpu.cmd` on Windows)

* Download [`yolov3-tiny.weights`](https://pjreddie.com/media/files/yolov3-tiny.weights) to the `bin` directory and run `./tiny-yolo.sh`

How to use **INT8**-inference:

* Use flag `-quantized` at the end of command, for example, [`tiny-yolo-int8.sh`](https://github.com/AlexeyAB/yolo2_light/blob/master/bin/tiny-yolo-int8.sh) or [`yolo_cpu_int8.cmd`](https://github.com/AlexeyAB/yolo2_light/blob/master/bin/yolo_cpu_int8.cmd)

* For the custom dataset, you should use `input_calibration=` parameter in your cfg-file, from the correspon cfg-file: [`yolov3-tiny.cfg`](https://github.com/AlexeyAB/yolo2_light/blob/29905072f194ee86fdeed6ff2d12fed818712411/bin/yolov3-tiny.cfg#L25) or [`yolov3.cfg`](https://github.com/AlexeyAB/yolo2_light/blob/29905072f194ee86fdeed6ff2d12fed818712411/bin/yolov3.cfg#L25), ...

How to use **BIT1-XNOR**-inference - only for custom models (you should train it by yourself):

* You should base your cfg-file on [`yolov3-spp_xnor_obj.cfg`](https://github.com/AlexeyAB/darknet/files/2853459/yolov3-spp_xnor_obj.cfg.txt) and train it by using this repository as usual https://github.com/AlexeyAB/darknet by using pre-trained file [`darknet53_448_xnor.conv.74`](https://drive.google.com/open?id=1IT-vvyxRLlxY5g9rJp_G2U3TXYphjBv8)

* Then use it for Detection-test or for getting Accuracy (mAP):

    * `./darknet detector test data/obj.names yolov3-spp_xnor_obj.cfg data/yolov3-spp_xnor_obj_5000.weights -thresh 0.15 dog.jpg`

	* `./darknet detector map data/obj.data yolov3-spp_xnor_obj.cfg data/yolov3-spp_xnor_obj_5000.weights -thresh 0.15`

Other models by the link: https://pjreddie.com/darknet/yolo/