Papers of the day   All papers

NEURAL TANGENTS: FAST AND EASY INFINITE NEURAL NETWORKS IN PYTHON

Comments

Dec 06 2019 Sam Schoenholz

After a ton of work by a bunch of people, we're releasing an entirely new Neural Tangents. Paper: https://arxiv.org/abs/1912.02803 Github: https://github.com/google/neural-tangents Colab Notebook: https://colab.sandbox.google.com/github/google/neural-tangents/blob/master/notebooks/neural_tangents_cookbook.ipynb
6 replies, 510 likes


Dec 06 2019 Jascha

Infinite width networks (NNGPs and NTKs) are the most promising lead for theoretical understanding in deep learning. But, running experiments with them currently resembles the dark age of ML research before ubiquitous automatic differentiation. Neural Tangents fixes that. https://twitter.com/sschoenholz/status/1202988151569973248
2 replies, 282 likes


Dec 07 2019 hardmaru

Neural Tangents is a Python library designed to enable research into “infinite-width” neural networks. They provide an API for specifying complex neural network architectures that can then be trained and evaluated in their infinite-width limit. 🙉🤯 https://arxiv.org/abs/1912.02803
1 replies, 130 likes


Dec 06 2019 Statistics Papers

Neural Tangents: Fast and Easy Infinite Neural Networks in Python. http://arxiv.org/abs/1912.02803
0 replies, 30 likes


Dec 06 2019 Jascha

Paper: https://arxiv.org/abs/1912.02803 Github: https://github.com/google/neural-tangents Colab Notebook: https://colab.research.google.com/github/google/neural-tangents/blob/master/notebooks/neural_tangents_cookbook.ipynb
0 replies, 28 likes


Dec 08 2019 Matthew McAteer

Easier “infinite-width” neural networks. 🤯 Seems like a pretty brilliant way to compress larger NNs. Plus, they've got a colab notebook demo showing how to build them in Jax (which I guess is what all the cool kids are using these days). https://colab.research.google.com/github/google/neural-tangents/blob/master/notebooks/neural_tangents_cookbook.ipynb https://t.co/1PFQx3wcht
0 replies, 18 likes


Dec 07 2019 MONTREAL.AI

Neural Tangents: Fast and Easy Infinite Neural Networks in Python Novak et al.: https://arxiv.org/abs/1912.02803 #DeepLearning #NeuralNetworks #Python https://t.co/JuzLIh5kxy
0 replies, 4 likes


Content