ML Model Tensorboard

mattmagic

Hey,

I am currently trying to run tensorboard in the visual analysis for my keras model.

Here are all packages installed in my conda env:

1. scikit-learn>=0.19.1,<0.20
2. statsmodels>=0.8,<0.9
3. jinja2>=2.10,<2.11
4. tensorflow
5. flask
6. tensorboard
7. keras
8. xgboost
9. h5py=2.7.1
10. pillow=5.1.0
11. matplotlib
12. gensim
13. nltk
14. seaborn
15. bokeh

When trying to switch to tensorboard I am getting the following exception:

Backend died before start, caused by: CustomPythonKernelPythonException: Failed to run webapp backend

HTTP code: 500, type: com.dataiku.dip.webapps.backend.WebAppBackendRunner$BackendStartFailedException

Did anyone succeed running tensorboard on dataiku?

Cheers,

Matthew

jimher

Hi Matt,

for me Tensorboard doesn't work either on Python 3.6. Creating a Python 2.7 environment seemed to work for me.