Workshop: Introduction to PyTorch with Dask
In this hands-on workshop, attendees will learn how Dask and parallelization can be incorporated into standard PyTorch workflows to create faster inference and training, as well as higher quality training results. Instructors will walk step-by-step through how to run two types of computer vision jobs on GPU Dask clusters: large batch inference and transfer learning.
The workshop will take place within Saturn Cloud Hosted, a platform that enables quick and easy access to parallel computing in Python. Attendees will receive free credits to Saturn Cloud Hosted to facilitate learning during and after the session. Join us to get hands-on experience using Saturn Cloud and learn what you need to confidently scale up your own deep learning.
After this workshop you will know:
Essentials of how Dask works, and how to apply it
How to parallelize an inference task in PyTorch with Dask
How to use the newly released Dask-PyTorch Python library
How to make a transfer learning workflow parallel on a multi-worker Dask cluster
Attendees need not have prior knowledge of Dask, but prior experience with PyTorch or other deep learning frameworks in Python are highly recommended.
Try Saturn Cloud Hosted for free here: https://www.saturncloud.io/s/tryhoste...