Everything is working for me, but it takes more than 8GB of GPU. (nvidia-smi)
I encountered the same thing with Darknet and was able to reduce the memory use by reducing the resolution (width/height) of the network in the configuration file, is there a way I can do it here with TensorFlow?
if not is there a way to unload the network from memory in python and reload it again when needed? or any other solution?
Everything is working for me, but it takes more than 8GB of GPU. (nvidia-smi) I encountered the same thing with Darknet and was able to reduce the memory use by reducing the resolution (width/height) of the network in the configuration file, is there a way I can do it here with TensorFlow?
if not is there a way to unload the network from memory in python and reload it again when needed? or any other solution?
thanks!