search

GoogleCloudPlatform / tensorflow-without-a-phd

A crash course in six episodes for software developers who want to become machine learning practitioners.
Apache License 2.0
2.78k stars 907 forks source link

Better explanation of how running job on GCP (Google Cloud Platform) #14

Open ClaudeCoulombe opened 6 years ago

ClaudeCoulombe commented 6 years ago

Greetings folks, I've a small suggestion to improve the explanations about running a model on the Google Cloud Platform, (section Run training job in the README.md) for the newbies. Do whatever you want with it...

In order to submit the training job to the Google Cloud Machine Learning Engine from your laptop:

1) Install the GitHub tensorflow-rl-pong code and requirements on your laptop 2) Get a Google Cloud Account and enable billing (credit card) 3) Create a Google Cloud Platform project on Google Cloud 4) Install the GCP command line tools on your laptop 5) Create a Google Cloud Data bucket in order to store the code and the trained model on Google Cloud 6) On your laptop, you should edit the submit.sh and change the commented line
# GCS_BUCKET="gs://my-bucket/" and replace my-bucket by your bucket id you've created in 5) 7) On your laptop, do bash submit.sh or eventually sudo bash submit.sh

ClaudeCoulombe commented 6 years ago

By the way, I've shared your excellent tutorial, video and GitHub code repo with the G+ Deep Learning and G+ AI Machine & Deep Learning (moderated) that are surprisingly still very active.

martin-gorner commented 6 years ago

Thank you for the suggestion. I will improve the instructions. And thanks for sharing.