exception during SEAM inference

YudeWang / SEAM

Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentation, CVPR 2020 (Oral)

MIT License

539 stars 97 forks source link

exception during SEAM inference #18

Open saysth opened 3 years ago

saysth commented 3 years ago

Dear YudeWang, I have successfully trained the model but during the stage of SEAM inference, the code will stop at a random iter(38, 14 or whatever) and don't go on and out_cam(or out_crf) folder won't produce file anymore. How can I solve the exception?

SCoulY commented 3 years ago

@saysth Hi, did u train ur model from scratch? Can u share ur hyperparam setups? I encounter problems in training from scratch using their default setup.

saysth commented 3 years ago

@SCoulY I just followed the instruction 1. SEAM training: python train_SEAM.py --voc12_root VOC2012 --weights ./xxx.pth --session_name $your_session_name I seemed to modify the line91 as weights_dict = torch.load(args.weights, map_location='cpu')

li-shuang1997 commented 3 years ago

@saysth Hi,I also encountered the same problem as you. Have you solved it? How was it resolved?

saysth commented 3 years ago

@li-shuang1997 no, i haven't

Fennyhhh commented 3 years ago

I think it may because line68: thread_pool = pyutils.BatchThreader(_work, list(enumerate(img_list)), batch_size=12, prefetch_size=0, processes=args.num_workers) cam_list = thread_pool.pop_results() here, the code use threads . it may cause this problem

Fennyhhh commented 3 years ago

I have changed it as: cam_list = [ _work(i[0],i[1]) for i in list(enumerate(img_list))] I hope it may help you.

@li-shuang1997 no, i haven't