TiagoCortinhal / SalsaNext

Uncertainty-aware Semantic Segmentation of LiDAR Point Clouds for Autonomous Driving
MIT License
417 stars 102 forks source link

acc on the nuscenes mini dataset #80

Open ZT-AIA opened 8 months ago

ZT-AIA commented 8 months ago

Hi, due to arithmetic limitations I am only training and validating on the nuscenes mini dataset, trained 3 times each 400 rounds the acc is 44.3%, 46.9%, 45.1% respectively, is this a normal level, have you done the same here?