improve baseline model training - Githubissues

DianaDI / mde

GNU General Public License v3.0

0 stars 0 forks source link

improve baseline model training #42

Closed DianaDI closed 4 years ago

DianaDI commented 4 years ago

add edge-based weighted l1-loss component (laplacian, canny)
add interpolation on ground truth
add more visualisations while training