Closed simonm3 closed 7 years ago
@simonm3 code should be running correctly. Can you give me some details?
keras-tqdm
or the development branchTry jupyter notebook
and the notebook in the examples
directory. Both the fit
and fit_generator
examples should be showing 2 progress bars, 1 for training and 1 for the epoch. Please let me know if the examples work correctly for you, or if your custom code is having issues.
Cheers
keras 1.1.2 python 3.6 jupyter pypi keras-tqdm
I get the two progress bars. But they only update once at the end of each epoch. And if there are 6 epochs they all say 6/6 at the end. I will try the examples tomorrow but don't think am doing anything especially weird.
On 8 April 2017 at 20:39, Ben notifications@github.com wrote:
@simonm3 https://github.com/simonm3 code should be running correctly. Can you give me some details?
- Keras 1.2 or 2
- Python 2 or 3
- console or Jupyter
- PyPI keras-tqdm or the development branch
Try jupyter notebook and the notebook in the examples directory. Both the fit and fit_generator examples should be showing 2 progress bars, 1 for training and 1 for the epoch. Please let me know if the examples work correctly for you, or if your custom code is having issues.
Cheers
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/bstriner/keras-tqdm/issues/11#issuecomment-292740856, or mute the thread https://github.com/notifications/unsubscribe-auth/ABJN6a4cCY2zGPyvqB-l6RtL0xA_xyY-ks5rt-J7gaJpZM4Mwl77 .
I'll give a quick try with keras 1.1.2 on py 2.7.
@simonm3 thanks for the find! Samples and epochs were swapped for Keras 1. If you install the current git master it should work for you. Please let me know and I can release the fix to PyPI.
@simonm3 just dropped keras-tqdm 2.0.1 on PyPI. If a pip install keras-tqdm --upgrade
fixes your problem please close the issue.
Cheers
Perfect thanks.
Great package. really useful. however does not update the epoch bar when using fit_generator. Looks like the code has started to be adapted for that but is not quite finished? Is that right?
Can I just lower the number of samples per epoch to get more frequent updates? Or will that have side effects?