Bugs in training

[ys-zong]

Hello, when I tried to train the model, there are several attributes not defined in the code.

1. self.loss_boxes

[rank3]:   File "/path/InstructSeg/instructseg/model/mask_decoder/mask_criterion/M
ask_Criterion.py", line 280, in forward                                                             
[rank3]:     losses.update(self.get_loss(loss, outputs, targets, indices, num_masks))               
[rank3]:   File "/efs/user/yongshuo/proj/InstructSeg/instructseg/model/mask_decoder/mask_criterion/M
ask_Criterion.py", line 245, in get_loss                                                            
[rank3]:     'boxes': self.loss_boxes,                                                              
[rank3]:   File "/opt/conda/envs/instr/lib/python3.10/site-packages/torch/nn/modules/module.py", lin
e 1729, in __getattr__                                                                                 
[rank3]:     raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")      
[rank3]: AttributeError: 'InstructSeg_criterion' object has no attribute 'loss_boxes'. Did you mean:
 'loss_labels'?

2. self.generative_model here.

Could you kindly advise how to fix them? Thanks!

[congvvc]

Thank you for your interest in our work. We have updated the code and resolve the above issues.
You can try the training stage again.