Computer Vision with Image Preprocessing and Tensor Processing Units (TPU)

In this video, two of my students at Washington University in St. Louis, Lyndon He and Tony Liu describes their class project that took first place of 30 teams in my course. They made use of Tensor Processing Units (TPU) to speed their process, as well as image preprocessing to detect if block structures were stable or unstable. This team made 54 submissions to the Kaggle In Class competition, great job Lyndon and Tony!

Guest Speakers: Lyndon He and Tony Liu

Kaggle Competition:

Code Available Here:

[rt_shll]

Killing it Jeff !! Love these videos. One of the best ML channels by far. You should run an online course - I am sure lots of folks would join. I wish I was closer to take your course. By far one of the best styles of teaching, love the open nature of how your share code, notebooks and help other grow in this space. Kudos Jeff,

[patrikpatrik]

Hej Jeff, thanks for your videos, I have learnt so much from you!!! I have one suggestion for a video. I have been using colab for training, but lately I have found the stability to be questionable as well as the time between I update a file in gdrive until colab accept the updated file to be too long, so I am concidering changeing from colab to AWS. Would it be possible for you to make a video of how to get tensorflow up and running in AWS?