Machine Learning with TensorFlow
summary
TensorFlow is an open C++ core with Python and other front ends that runs on CPU GPU and TPU and claims 1,000,000+ downloads and nearly 500 contributors (over half outside Google) after one year. It expresses models as dataflow graphs, automatically parallelizes on-device ops, and shards cross-device edges with send/receive nodes to scale; users report 56x speedup over 100 machines for synchronous SGD. Concrete builds include a Raspberry Pi cucumber sorter and hobby cat-detecting sprinkler integrations that work in practice.
AI-written summary · about
chapters (21)
- 0:00Intro
- 1:21Neural networks
- 1:41Inception (2015)
- 2:03That's not all
- 4:51Portability is a requirement
- 6:53machine learning gets complex quickly
- 10:01TensorFlow powered Cucumber Sorter
- 10:57Platforms
- 12:22TensorFlow Architecture
- 13:21Ecosystem
- 13:41Parallelism in Op implementations
- 14:36Task Parallelism in DataFlow graph
- 15:02Data Parallelism
- 15:17Model Parallelism
- 15:46Distribution across Devices
- 17:39Scale across machines
- 19:28Align cognitive model with programming model
- 20:00TensorFlow contains complete algorithms
- 20:42Simple machine learning
- 21:49TensorFlow handles complexity for you...
- 22:21Ready to use Machine Leaming models
related talks
13mTaking Machine Learning to the Next Level
07mMachine learning tools from Google
28mTensorFlow 1.0: open source machine learning
39mSwift for TensorFlow: Machine Learning with No Boundaries
33mWhat's new in Machine Learning
31mTensorflow for Janitors
54mAdventures in Javascript Machine Learning
06mEngage with the Machine Learning community