liyang
liyang
Hi, @maria8899, @kilsenp, Have you reproduced the results on MultiMNIST or CityScapes? Thanks.
@youandeme, I haven't reproduced the results on MultiMNIST. I used the same hyper parameters mentioned by the author in https://github.com/intel-isl/MultiObjectiveOptimization/issues/9, but can not surpass the uniform scaling baseline. Also, I...
Hi, @ozansener, as you mentioned, the order of different methods (MGDA-UB vs. uniform scaling) will keep the same whatever test set I use. But on the validation set, I cannot...
Thanks. @ozansener. On CityScapes I re-run your code with instance & semantic segmentation task and get the following results for MGDA-UB and SINGLE task, respectively: | method | instance |...
@ozansener, I use exactly the hyper-params you posted for single & multi-task training. I use 1/0 and 0/1 scale for single task training (instance and semantic segmentation) and didn't do...
Hi, @bsaint, @milos-popovic, Have you figured out the discrepancy and reproduced the results on MultiMNIST, CelebA or CityScapes? Thank you.
The loss is as usual, I think you can use the normal mask loss such as BCE.
Have you checked whether all the layers that are supposed to be in one group are gathered correctly by the algorithm?
@jwyang @doronpor @Feynman27 Hi, have you guys fix the bug? Thanks.