YOLOV2 training using GPU

Hi

Trying to run the following code on GPUCoder and error that trainYOLOv2ObjectDetector is not supported.

[detector,info] = trainYOLOv2ObjectDetector(trainingData,lgraph,options);

If the above code is not supported by the GPU coder, what are the alternatives to accelerate YOLOV2 training process?

Regards

Answer 1

[Shivam Singh]

Hello gil,

Code generation using GPU Coder™ for trainYOLOv2ObjectDetector is not supported.

Yolov2 training process can be accelerated by using GPUs.

[detector,info] = trainYOLOv2ObjectDetector(trainingData,lgraph,options);

To access the GPUs, you can set the “options” in the trainYOLOv2ObjectDetector function, by referring the ExecutionEnvironment in the trainingOptions function.

Once the detector is trained and evaluated, then you can generate code for the yolov2ObjectDetector using GPU Coder™. For more information related to Code generation using GPU Coder™, you can refer the following:

• Object Detection Using YOLO v2 Deep Learning
• Code Generation for Object Detection by Using YOLO v2