Running the example on Google Cloud Datalab about model-analysis HOT 5 CLOSED

I was able to make some progress. Currently executing the Chicago taxi on the Google Cloud Datalab. I was able to import the tensorflow-model-analysis, tensorflow-data-validation, render the data validation tables. The only piece that is currently not working is the tensorflow-model-analysis rendering commands. See attached the notebook (with txt extension to allow me to uplaod) Please let me know if you have any recommendation. Many thanks, eilalan

or_chicago_taxi_tfma_local_playground.ipynb.txt

from model-analysis.

@OrielResearchCure ,
Can you please confirm if you can run Tensorflow Model Analysis rendering commands with the latest version.

from model-analysis.

Hi,
Thank you for the follow up. Datalab is my daily go to tool.

As far as I remember, It is not working smoothly.
I was running the below installations when I wanted it to work.
TFX is not supporting tensorflow-alpha at the moment, so I am unable to work with it at all at the moment. I am running the same list of installations (changing tensorflow==2.0.0-alpha0) when I want to work with the latest tensorflow.

I am looking forward to have the latest libraries working out of the box.
Please let me know if there is a way to work with TFX / latest Tesnorflow with less installations added.

Many thanks,
Eila

%%bash
source activate py2env
conda install -y dill==0.2.8.2 psutil==5.4.8 protobuf==3.6.0
conda remove -y python-snappy
pip install tensorflow==1.12.0 tensorflow-model-analysis==0.11.0 tensorflow_transform==0.11.0 pandas-gbq==0.4.1 tensorboard==1.12.0 apache-beam[gcp]==2.8.0 tensorflow-data-validation==0.11.0 tensorflow-probability tensorflow-metadata==0.9.0 widgetsnbextension==3.4.2 scipy==1.1.0 tensorflow-serving-api==1.11.0

from model-analysis.

@OrielResearchCure ,

1. Supporting TF 2.0 end-to-end in TFX is currently a work in progress.
2. As of now, installations are required because each Component can be used independently as well.

from model-analysis.

Automatically closing due to lack of recent activity. Please update the issue when new information becomes available, and we will reopen the issue. Thanks!

from model-analysis.