last fun thing to think about is that we're doing 1.28M images over 90 epochs with 68K batches, so the entire optimization is ~1700 updates to converge. How lucky for us that our Universe allows us to trade that much serial compute for parallel compute in training neural nets

NeverGrad: gradient free optimization library. Now open source. code.fb.com/ai-research/ne…

Impressed by DigitalOcean and their ability to scale Docker with Kubernetes with developer friendly interface

Consider joining one of the leading AI groups in the world

One of the brightest minds in AI has launched Primer

Thank you for providing this exceptional resource to the #AI community

Extraordinary progress on #GAN with BigGan creating novel scenes based on input from Planet Earth footage

Excited to try out the newly open sourced Flashlight (C++ ML framework) and Wav2letter++ (rapid speech to text) from FB. See code.fb.com/ai-research/wa…

Another impressive text to speech implementation able to carry out numerous tasks including discerning different speakers.

Our model allows for data to stay secure within an organization but the resultant trained models shared widely across the community for assessment.

Check out Christian Guttmann excellent #AI presentations about benefits and trends.

New NLP work by Google AI and CMU School of Computer Science called Transformer-XL now on GitHub

Excellent progress by the OpenAI team in building and training unsupervised language models