EOFError: Ran out of input on windows when training with CSV

[devedse]

I'm currently running into an issue when I run the following command:

retinanet-train C:/XGitPrivate/*****/ImagesCarsExport/annotations.csv C:/XGitPrivate/*****/ImagesCarsExport/classes.csv

The logs of this command:

classification (Concatenate)    (None, None, 1)      0           classification_submodel[1][0]
                                                                 classification_submodel[2][0]
                                                                 classification_submodel[3][0]
                                                                 classification_submodel[4][0]
                                                                 classification_submodel[5][0]
==================================================================================================
Total params: 36,382,957
Trainable params: 36,276,717
Non-trainable params: 106,240
__________________________________________________________________________________________________
None
Epoch 1/50
Exception in thread Thread-2:
Traceback (most recent call last):
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\threading.py", line 916, in _bootstrap_inner
    self.run()
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\threading.py", line 864, in run
    self._target(*self._args, **self._kwargs)
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\site-packages\keras\utils\data_utils.py", line 565, in _run
    with closing(self.executor_fn(_SHARED_SEQUENCES)) as executor:
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\site-packages\keras\utils\data_utils.py", line 548, in <lambda>
    initargs=(seqs,))
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\context.py", line 119, in Pool
    context=self.get_context())
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\pool.py", line 174, in __init__
    self._repopulate_pool()
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\pool.py", line 239, in _repopulate_pool
    w.start()
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\process.py", line 105, in start
    self._popen = self._Popen(self)
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\context.py", line 322, in _Popen
    return Popen(process_obj)
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\popen_spawn_win32.py", line 65, in __init__
    reduction.dump(process_obj, to_child)
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
TypeError: can't pickle generator objects

Using TensorFlow backend.
Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\spawn.py", line 105, in spawn_main
    exitcode = _main(fd)
  File "C:\Users\****\AppData\Local\conda\conda\envs\POMP\lib\multiprocessing\spawn.py", line 115, in _main
    self = reduction.pickle.load(from_parent)
EOFError: Ran out of input

I already found a solution here: #857 that you should add --workers 0.

However since this is not really a 'solution' and is more of a workaround I would like to leave this issue open to see if this can be fixed at the underlying level.

[hgaiser]

However since this is not really a 'solution' and is more of a workaround I would like to leave this issue open to see if this can be fixed at the underlying level.

The underlying issue comes from Keras it seems; not much we can do about that in this repository.

[devedse]

Would you be able to help produce a minimal sample to reproduce this issue so we could pass it on to the Keras team?

[hgaiser]

I'm not using windows (it seems to only be an issue on windows); I think you're better equipped to work out a minimal example.

[hgaiser]

Closing this as it seems there is no issue for keras-retinanet.