File create_strand_latent_weights.py

utils/create_strand_latent_weights.py:None–None

#!/usr/bin/env python3

#when training the diffusion model we apply a L2 loss on all the channels of the scalp texture. However some of the channels are essentially noise. 
#This is due to the fact that some dimensions of the strand vae encode very little information and don't signficantly modify the strand shape.