AssertionError

[liunian199]

when i run，it goes by init stage4.2.fuse_layers,then assertion error occurs

[liunian199]

complete sentence can be expressed as"assert(isinstance(orig, torch.nn.Module))"

[liunian199]

i run it with only one GPU

[mytk2012]

i have succeeded，thanks

[Janisco-cv]

I have also encountered this problem.I tried some methods but they didn't work out. Do you have any successful suggestions?

[zeven11]

I have encountered this problem too.first the line is "args.world_size = ngpus_per_node * args.world_size"in File "tools/dist_train.py", line 108,so i think the problem is about MULTIPROCESSING_DISTRIBUTED,so i change file config _C.MULTIPROCESSING_DISTRIBUTED to False,the error is gone,but there is a new error"File "tools/dist_train.py", line 196, in main_worker writer_dict['writer'].add_graph(model, (dump_input, ))" ,I think this problem is about tensorboardx.

[DuckJ]

@zeven11 Have you solved your problem? I have the same error writer_dict['writer'].add_graph(model, (dump_input, ))

[ANGDL]

Layers include None, so assert failed. I commented out this line.

[HoBeom]

+1 Same error

[annie-surla]

Layers include None, so assert failed. I commented out this line.

Can you mention what lines of the code you commneted out in the model file?

[madokainoue1173]

I have encounterd this problem, too. I modified tensorboardX version 2.1 to 0.8 as below. pip install tensorboardX==0.8

[annie-surla]

@madokainoue1173 Actually i tried with torch==1.5 and tensorboardX==2.0 and it worked for me.