I’m trying to figure out why Tensorflow can only obtain 9 of the 11gb available? I assumed since it has the same architecture as the titan xp but just 1 less gb of ram, it would be a killer deal for deep learning. What’s the deal?
how are you running the container? Is maybe the host system using your missing memory?
This forum is for users of the NVIDIA GPU Cloud. NGC isn’t supported on the GTX 1080 Ti so this forum isn’t focused on that platform.