Best practice to run Tensorflow app on GCP? - tensorflow

I wish to run a Python app that uses Tensorflow to run simulations and outputs results to a csv file.
I want to run it on GCP. My plan was to run it in a Dataproc cluster using Tony.
It seems there are many GCP ways of doing ML stuff such as using AI Platform, and I wondered if there are easier/better ways of achieving my aim.

I would suggest to use Google Cloud AI platform to achieve your goal, because if you do not have dependency on Hadoop ecosystem there no need to use Tony on Dataproc and it should be much easier to use Google Cloud AI platform for your use case.

Related

How to open Google Colab files in Neovim?

I am looking for a way to open Google Colab files in Neovim. I am basically trying to figure out a way to connect the two platforms since I have to work on Google Colab for some CUDA related assignment stuff and I just love Neovim, By doing this, I would also be able to use Neovim cheatsheets, Github Copilot and other productivity tools.
Any advice that would get me a step closer would be highly appreciated

Deploying Jupyter notebooks as a component in AI Platform pipelines

I have a Jupyter Notebook containing the model creation, model deployment on the AI Platform, and creating versions. I am able to get the predictions for my model. Now I am trying to build a CI/CD pipeline automating the entire process. Is there a way that I can pass in my entire Jupyter notebook as a component in AI Platform Pipelines?
You can use Papermill to create a parameterized notebook, which can then be executed via CI/CD. This article explains a bit more in detail in the 'Reproducible Notebooks' section.

Tensorflow Federated TFF still a Simulation Environment?

TensorFlow Federated (TFF) is an open-source framework for ML and other computations on decentralized data.
As per Stack overflow link
TFF only provides a simulation environment for use in Federated
Learning (FL) research. There is not yet a "real world" FL deployment
platform.
But, tensorFlow release history shows that now there are many release versions for TF 2.x as well.
https://github.com/tensorflow/federated/releases
Can anybody comment, if TFF is still or simulation environment or can be used as "real world" FL deployment platform?
At this time, TensorFlow Federated does not have out-of-the-box support for what would generally be considered production federated learning. A production-level deployment runtime still needs to be built.
For different flavors of federated learning this maybe easier or harder.
It may be possible to create a system for cross-silo FL using the executor components already available in TensorFlow Federated. Particularly it maybe possible to extend and build something on top of the remote execution stack (e.g. tff.framework.RemoteExecutor)
However for cross-device FL it maybe signifcantly more work, as there are no integrations or components for deploying and execution computations on mobile operating systems (iOS, Android, etc) yet.

Export a python dataframe from Google AI Platform Notebook to a table in Google BigQuery

I got a dataframe in Google AI Platform Notebook (PN) that i would like to transfer to a table in Google BigQuery.
I am aware of the option to use: df.to_gqb() but that requires a pip install of pandas_gbq. I prefer to avoid pip installs on top of the libraries already included in PN to keep the setup as simple as possible.
Do i perhaps miss an easy solution?
Br, Torben
There is a native library for importing data into BigQuery, via the BigQuery Client Library.
As the "google.cloud" library is already part of AI Platform notebooks environment, you won't need to install any new packages (thanks Torben for the confirmation!)
Here is a link to the official documentation

CNTK deployment for real time predictions

TensorFlow has a separate project for its production usage, as noted here, called TensorFlow Serving.
How should I use CNTK in a production environment, and how should I handle it's deployment? Hopefully one could deploy trained models in to a server/cluster that serves the model with RPC or HTTP REST API.
If no such tool exists, what should be the first steps to develop it and a good architecture to roll out on my own?
We do support serving CNTK models in a production environment. You can find information about model evaluation/inference: https://github.com/Microsoft/CNTK/wiki/CNTK-Evaluation-Overview. For deployment, you need deploy the dlls specified here. A tutorial for deploying CNTK in Azure is available here.
No such tool exists from the CNTK team, but the new APIs in C++ or python should make the task pretty easy once you have a trained model.