I would like to run some operations (sigmoid...)of a custom YOLOV4 tiny model (size 512) that are not supported by Tensil framework on CPU. To do this, I need to generate a TF lite model by changing outputs of the last convolution layers to be inputs. Any suggestions on how to do it?
Hi,
I would like to run some operations (sigmoid...)of a custom YOLOV4 tiny model (size 512) that are not supported by Tensil framework on CPU. To do this, I need to generate a TF lite model by changing outputs of the last convolution layers to be inputs. Any suggestions on how to do it?
Thanks.