High ram usage with tlt unet training

[Morganh]

Sure, you can, let us sync in topic ValueError: No such layer: block_1b_relu in unet nvidia-tlt 0.1.4 - #3 by ksokolov

If no further question, we can close this topic.