Exception has occurred: ValueError - Error during the execution of train.py script

[KevinSIERA]

Hello I tried to reproduce the result of this method and got this bug :
Exception has occurred: ValueError
No variables provided.
File "/home/kevin/3DSSD/lib/core/trainer.py", line 87, in __init__
self.update_op = [self.optimizer.apply_gradients(zip(self.grads, self.params), global_step=self.global_step)]
File "/home/kevin/3DSSD/lib/core/trainer.py", line 216, in <module>
cur_trainer = trainer(args)

I'm on Ubuntu16.04, using python3.6 with tensorflow1.4. I also would like adding, I added these 2 lines in _build_model_list function to overcome an UnboundLocalError.
params = []
extra_update_ops = []

When I search to found this exception, I notice that self.params and self.extra_update_ops are respectively linked with cfg.TRAIN.CONFIG.TRAIN_PARAM_PREFIX and cfg.TRAIN.CONFIG.TRAIN_LOSS_PREFIX. These 2 parameters are not indicated in the 3dssd.yaml file. I tried to add them but I have no idea of their value and nothing changed when I add them ithout value. So how can I overcome this exception ?

[Antonskoltech]

@KevinSIERA did you fixed? The same problem