Bashintosh

see my comment [here](https://github.com/rbgirshick/py-faster-rcnn/issues/787#issuecomment-371476349)

I got the same issue when using a min_size to big according to my frame input size. Reducing the min_size to 1 (in my example) solved the problem

Here, someone has implemented yolov3 in tensorflow. https://github.com/mystic123/tensorflow-yolo-v3

Hi @ambr89, Thanks for the link, I will try it. It would be nicer if Yolov3 was supported in darkflow...

Great ! Which layer is not easy ? You may need some help ?

Do you think it will be add in darflow ? Actually, I don't understand why yolov3 could not be supported directly in darkflow ?

Maybe the support of tiny yolo v3 would be a good beginning. The shortcut layer is not used in Yolov3-tiny. So, supporting yolov3 tiny is only question of upsampling layer...

This bug is due to the definition of the function end into my linker script. I think that this name is used to manage the head. So We should rename...

for sure, here it is : ``` /**************************************************************************/ /* Configure memory regions */ MEMORY { VECTORS (rx) : ORIGIN = 0x0, LENGTH = 0x00c0 FLASHCFG (rx) : ORIGIN = 0x00000400,...

Thanks for the interest you have in telemetry. I think contributions would a great idea. > Le 6 juin 2017 à 18:08, Andrew a écrit : > > I'm looking...