While training with CPU on Windows, Python crashes

I’m trying to train a net with DLC on Windows, CPU. After 2000+ iterations, Python crashes with the error message “Python has stopped working”. Shortly after, Windows crashes as well O.o
Does anyone know why? I have 32GB RAM installed and the video I’m analyzing is only 640x480 (34mins though - is that a problem?).

Training only uses images that are in the labeled/data folder, so only what you labeled.

Can you tell us:

What version of tensorflow and ipython you have installed?

Can you post the full call you used, ie deeplabcut.train_network( …, )

( You may want to use COLAB anyhow to have access to a gpu. )

Python 3.6.9 |Anaconda, Inc.| (default, Jul 30 2019, 14:00:49) [MSC v.1915 64 bit (AMD64)]
Type 'copyright', 'credits' or 'license' for more information
IPython 7.8.0 -- An enhanced Interactive Python. Type '?' for help.
import tensorflow
tensorflow.__version__
Out[2]: '1.13.1'

My call is basic:

import deeplabcut
config_path = 'C:\\Users\\xxx\\test3-xxx-2019-11-11\\config.yaml'
deeplabcut.create_training_dataset(config_path)
deeplabcut.train_network(config_path)

I can’t use COLAB since I’m evaluating privacy-sensitive research data. I’ll hopefully soon have access to a decent NVIDIA GPU though.