Set up cloud-GPU ML model training on Google Colab
No fancy GPU-processor? No time to leave your machine crunching for endless hours on a single machine learning task? Problems installing the endless dependencies your model requires? No problem.
Colab Notebooks are Jupyter Notebooks that run in the browser, using $10/month cloud-GPU infrastructure from Google.
They're cheap
Training a style transfer model with satisfactory results might typically
cost between $15-$20 on Cloud GPU architecture for a single training. Colab is $10 a month
for all the training you can squeeze in.
They're easier
Colab machines are fairly optimized for Machine Learning tasks. They make
it easy to:
- install the libraries you need
- visualize the steps of your process through the notebook interface
- run a model training instance without bogging down the processor on your personal machine.
- There's lots of other Colab notebooks out there to modify for your purposes.
To get started, visit this link. The Colab notebook itself will help you work through your first style transfer model training.
