Shuo Wang
Shuo Wang
I'm also very confused about the implementation of importance calculation which is obtained by the input dimension.
@666zz666 Don't you think that there might exist a bug in main.py at line mp.spawn and main_worker accept 3 parameters?
In our work, HEAD is just DCP. The backbone is defined in resnet_def/*.py. We only use backbone to extract feature which is later used to do matrix product with dcp.
@khawar512 You can simply rename lru_python_impl.py to lru_utils.py so you can fix ModuleNotFoundError. For ArcMarginProduct, member function add_margin in ffc_ddp.py plays the same role as ArcMarginProduct. At line 34 in...
@khawar512 FFC is not a loss layer. It's proposed to deal with the case where id is extremely huge and cannot be fed into a single gpu while distributed partial...
@khawar512 net type is the backbone you use like r50, r100, efficient net and other type of networks etc.
@xiakj world_size can be arbitrary positive number. I found you may make a mistake. World_size is a positional argument not optional argument.
wh > mobile_ms_0.01id The error is obvious. It takes "mobile_ms_0.01id" as int value. I guess you may place arguments in wrong place.
Thanks for your attention to our work. Well, in our experiments, we use LFW as our validation set.