directory existance issues

trying to train my first model. using png images and all default settings. I get the error **FileExistsError: [Errno 17] File exists: 'plots'** okay… so let’s remove that folder. same error. Here’s the full trace:

File "/home/xxx/anaconda3/envs/imclass/bin/transfer", line 11, in <module>
    load_entry_point('transfer==0.36', 'console_scripts', 'transfer')()
  File "/home/xxx/anaconda3/envs/imclass/lib/python3.6/site-packages/transfer-0.36-py3.6.egg/transfer/__main__.py", line 123, in main
    project = train_model(project, final = args.final, last = args.last)
  File "/home/xxx/anaconda3/envs/imclass/lib/python3.6/site-packages/transfer-0.36-py3.6.egg/transfer/model.py", line 75, in train_model
    os.makedirs(plot_path)
  File "/home/xxx/anaconda3/envs/imclass/lib/python3.6/os.py", line 220, in makedirs
    mkdir(name, mode)
FileExistsError: [Errno 17] File exists: 'plots'

This also prints out when it begins to run, but I don’t think it’s related at all:

lib/python3.6/site-packages/h5py-2.7.1-py3.6-linux-x86_64.egg/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters

though it seems like a reasonable thing to fix.

[matthew-sochor-zz]

@mpilosov I think I fixed your issue in https://github.com/matthew-sochor/transfer/commit/804487a22d89c35565c1f2238664313d1335f5d8

I pushed up to version 0.37 on PyPi, please give it another try and let me know if this resolves things for you. And again, thanks for the issue and I hope this works for you.

[mpilosov]

I’m installing from source using python setup.py install using a fresh conda environment and would mention that build fails because flask isn’t in your dependency list.

having path-completion in the setup would help as well, I did have to constantly check the configure file to ensure everything was entered right (found a typo, fixed it), now it is finally running. I see a lot of val_categorical_accuracy did not improve so I’m not sure how well it is going, and am excited to see the end result…

thanks for the fix! hope this turns into something useful. I’m training it on some synthetic image data that should be very easy for it to learn. starting with a small batch (likely too small) just to make sure the algorithm completes. If I add more images to the directories, simply doing transfer --run will pick up where it left off or restart the training?