search

xingyul / flownet3d

FlowNet3D: Learning Scene Flow in 3D Point Clouds (CVPR 2019)
MIT License
359 stars 83 forks source link

optimization loss #40

Closed Yang-L1 closed 4 years ago

Yang-L1 commented 4 years ago

Hi, Line 119 in train.py train_op = optimizer.minimize(loss, global_step=batch) Should this loss be loss[0]? Since the loss is a tuple object from the get_loss() function.

Btw, was the pretrained model trained by a huber loss or L2 loss?

xingyul commented 4 years ago

loss is a scalar tensor. The pretrained model is trained under L2 loss.