search

Sense-GVT / Fast-BEV

Fast-BEV: A Fast and Strong Bird’s-Eye View Perception Baseline
Other
596 stars 91 forks source link

The loss does not decrease during the training process #27

Open pengcheng001 opened 1 year ago

pengcheng001 commented 1 year ago

image

ymlab commented 1 year ago

Maybe run the evaluation to see if there is a reasonable accuracy. If not, see if the pretrained model and learning rate need to be adjusted.