Tensor Processing Units

Content collection

December 15, 2020

Tensor Processing Units (TPUs) are deep-learning specialized hardware accelerators available through Kaggle and Google Colab.

Getting Started Notebooks

Getting Started: TPUs + Cassava Leaf Disease is a Kaggle notebook built for anyone interested in creating a baseline model using Tensor Processing Units (TPUs) and begin making submissions to the Cassava Leaf Disease Classification competition. If you’ve taken the Kaggle Intro to Deep Learning and//or the Kaggle Computer Vision course you’ll find this notebook to be a good starting place to bridge what you’ve learned in our micro-courses and applying that knowledge to get started in a competition.

Getting Started: Rainforest EDA with TFRecords is a Kaggle Notebook that walks through an introductory-level exploratory data analysis (EDA) using the TensorFlow Simple audio recognition: Recognizing keywords tutorial and applying those learnings to the TFRecords that are available in the Rainforest Connection Species Audio Detection competition.

TPU Videos

TPUs, systolic arrays, and bfloat16 multipliers: accelerate your deep learning

Learn With Me: Getting Started with Tensor Processing Units

Learn With Me: Troubleshooting and Optimizing for TPUs

Additional TPU videos I’ve produced:

  • TPU Notebook Walkthrough: Introduction to TFRecords with Ryan Holbrook [VIDEO]

  • Pneumonia Classification on X-rays using TensorFlow and TPUs with Amy Jang [VIDEO]

Posted on:
December 15, 2020
1 minute read, 199 words
Tensor Processing Units Python Kaggle
See Also: