_pickle.UnpicklingError: invalid load key, '\x0a'.

[guoguotingzhu]

I followed your guidance and met the problem below when I run the code :
"python /home/liujianguo/jiguo/mask-rcnn.pytorch/tools/train_net_step.py --dataset coco2017 --cfg /home/liujianguo/jiguo/mask-rcnn.pytorch/configs/baselines/e2e_mask_rcnn_R-50-C4_1x.yaml --bs 2 --iter_size 1"

Traceback (most recent call last):
File "/home/liujianguo/jiguo/mask-rcnn.pytorch/tools/train_net_step.py", line 454, in
main()
File "/home/liujianguo/jiguo/mask-rcnn.pytorch/tools/train_net_step.py", line 256, in main
maskRCNN = Generalized_RCNN()
File "/home/liujianguo/jiguo/mask-rcnn.pytorch/lib/modeling/model_builder.py", line 124, in init
self._init_modules()
File "/home/liujianguo/jiguo/mask-rcnn.pytorch/lib/modeling/model_builder.py", line 128, in _init_modules
resnet_utils.load_pretrained_imagenet_weights(self)
File "/home/liujianguo/jiguo/mask-rcnn.pytorch/lib/utils/resnet_weights_helper.py", line 31, in load_pretrained_imagenet_weights
pretrianed_state_dict = convert_state_dict(torch.load(weights_file))
File "/home/liujianguo/anaconda3/envs/PANet/lib/python3.6/site-packages/torch/serialization.py", line 303, in load
return _load(f, map_location, pickle_module)
File "/home/liujianguo/anaconda3/envs/PANet/lib/python3.6/site-packages/torch/serialization.py", line 459, in _load
magic_number = pickle_module.load(f)
_pickle.UnpicklingError: invalid load key, '\x0a'.

my operating environment:
cuda 8.0
cudnn 7.3.0
pytorch0.4.0
python3.6
my GPU is geforce gtx 1070ti
I have no idea how to deal with it and really hope you can give me some suggestions.
I would appreciate it if you could reply to me .Best wishes for you.

[saadnajib]

did you find the solution ?

[Ankita-020696]

I am facing similar issue , any solution to fix it?