E0FError: Ran out of input, error training object detection

I am learning and following the seasons of videos of jetson fundamentals, but I got stacked here, at the moment of train object detection:

2021-05-21 08:35:54 - Init from pretrained ssd models/mobilenet-v1-ssd-mp-0_675.pth

Traceback (most recent call last):
File “train_ssd.py”, line 309, in
File “/jetson-inference/python/training/detection/ssd/vision/ssd/ssd.py”, line 119, in init_from_pretrained_ssd
state_dict = torch.load(model, map_location=lambda storage, loc: storage)
File “/usr/local/lib/python3.6/dist-packages/torch/serialization.py”, line 585, in load
return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
File “/usr/local/lib/python3.6/dist-packages/torch/serialization.py”, line 755, in _legacy_load
magic_number = pickle_module.load(f, **pickle_load_args)
EOFError: Ran out of input

I tried many times with different data and even reinstalled the jetson from scratch but I am still having the same error


The error is caused by the mobilenet-v1-ssd-mp-0_675.pth file.
Have you read it successfully on other environment?

Below is a relevant topic and please check if it helps first:


So , there is an empty file. Were you able to find it?
I get the same problem.

AttributeError: Can’t pickle local object ‘TrainAugmentation.init..’
(pytorch_train) PS C:\Users\thilo\jetson-inference\python\training\detection\ssd> 2021-07-28 14:09:29 - Using CUDA…
Traceback (most recent call last):
File “”, line 1, in
File “C:\Users\thilo\anaconda3\envs\pytorch_train\lib\multiprocessing\spawn.py”, line 116, in spawn_main
exitcode = _main(fd, parent_sentinel)
File “C:\Users\thilo\anaconda3\envs\pytorch_train\lib\multiprocessing\spawn.py”, line 126, in _main
self = reduction.pickle.load(from_parent)

SOLUTION: changed workers=1 to workers=0

Hi 79458,

Please help to open a new topic with more information. Thanks