What's new in TensorFlow for R

on YouTube ↗19m

" summary

TensorFlow 2.0 brings eager execution so code runs like regular R arrays instead of building a deferred graph, plus tf.autograph to turn R-style if/for into efficient TF ops. tfdatasets adds a spec interface for tabular data that learns vocabularies and scalers and plugs into Keras, and Keras now includes a text vectorization layer with adapt for deployment. tfhub models load as Keras layers and tfds provides easy access to large public datasets.

AI-written summary · about

" chapters (9)

1. 0:00Start
2. 0:27What's TensorFlow?
3. 3:39tfautograph
4. 4:23tfdatasets: Feature spec interface A recipes like interface for pre-processing tabular data for Deep
5. 6:22tfhub: R interface to TensorFlow Hub
6. 8:38keras: Text preprocessing layer
7. 10:22tfprobability
8. 11:33autokeras
9. 13:20Model packages

" related talks

• 32mGetting Started with TensorFlow 2.0Paige Bailey, Josh Gordon · Google I/O 2019
• 28mTensorFlow 1.0: open source machine learningAndrew Selle · FOSSASIA Summit 2017
• 33mWhat's new in Machine LearningKemal El Moujahid, Sarah Sirajuddin, Craig Wiley · Google I/O 2021
• 41mTensorFlow ever higherYoann Benoit · DevFest Paris 2019
• 47mSwifty TensorFlowSiamak Ashrafi · 360|iDev 2019
• 38mCutting Edge TensorFlow: New TechniquesElie Burzstein, Josh Dillon, Michael Bendersky, Sofien Bouaziz · Google I/O 2019
• 07mMachine learning tools from GoogleNadav Eiron · Google's Machine Learning Virtual Community Day 2021
• 05mWhat's New in the Torch EcosystemDaniel Falbel · posit::conf(2023)