Relative Content

Tag Archive for pythontensorflowkerasdeep-learning

Tensorflow training loop keeps appending to system memory over time

I am trying to use the encoder of a pre-trained UNet model to output features which are then used to train a densely connected MLP for a regression task. However, in the code shown below, there is some part that keeps appending to the system RAM over the course of 5-6 epochs, filling up all of the 54Gb of RAM of the machine.

PointNet DL model improvement, window segmentation

I’m making a project where I give a pointnet architekture a bunch of pointclouds from rooms and the points have indexes/labels. They re simple .xyz files but I add a forth row 1 or 0. 1 means it part of a window, 0 means its not.
So, the model making is running well with low loss and high accuracy. The accuracy increasing as it should.
But when I want to use it for segmentating it thinks ever point is part of a window on that pointcloud, its sure in it not even 0.99 threshold helps, it still exports the whole cloud. And I just can’t figure it out why is not working.
Here is the code witch makes the model: